GitHub Copilot logo
A

GitHub Copilot

A Tier · 8.3/10

**Nov 3-4: Copilot now works inside Xcode's own coding assistant (public preview from Xcode 27.3 beta, every Copilot plan) and inside JetBrains Air, where cloud tasks run on your Copilot plan rather than JetBrains AI credits. From Universe (Oct 28, recorded late): Copilot Memory is generally available on paid plans, and Copilot dreaming, which curates memories and proposes instruction-file changes as pull requests, is in public preview.** **Oct 28 (GitHub Universe): Copilot Fusion went GA on every Copilot plan, replacing auto model selection and HydraFusion with one router you run at Standard, High or Max effort (30% off model costs for a limited time), and GitHub Enterprise now includes Copilot for every licensed user for customers working with a GitHub or Microsoft account team, with no per-seat AI credits and usage metered against the enterprise budget; Copilot Standalone covers other users at $10 per user per month. The same day brought image generation and editing in Copilot (GPT Image 2.5 Flare, public preview), Projects and computer use in the Copilot app (public preview), a Max review effort and GA approvals in Copilot code review, Agent Merge on github.com pull requests, one cloud session that follows you across web, mobile, CLI, app, VS Code, Slack and Teams (GA), and 26 GitHub-verified connectors (public preview).** **GPT-6.1 Terra joined Copilot on 2026-10-26, OpenAI's launch day, for Pro, Pro+, Max, Business and Enterprise. The six-model removal executed on 10/19 as scheduled, and the global default policy for new features rolled out from 10/22 through 10/26.** **Oct 15: @GitHub in Slack and Microsoft Teams now works across multiple repositories in a single conversation (issues, pull requests, bug investigations), can be mentioned anywhere in a message, and can use shared Slack messages and threads as context.** **Gemini 4 Argon joined Copilot on 2026-10-13, the day Google made it GA, for Pro+, Max, Business and Enterprise at provider list price; on 10/15 GitHub scheduled Claude Opus 4.8, Claude Opus 5 and GPT-5.6 Terra for removal from every Copilot surface on 2026-11-16 (Opus 4.8 stays for individual annual subscribers), and HydraFusion gained a Max option and a pool of Opus 5.5, GPT-6 Astra, GPT-6.1 Sol and GPT-6 Luna (10/14).** **Claude Haiku 5.5 joined Copilot on 2026-10-07, the day Anthropic launched it, for Pro, Pro+, Max, Business and Enterprise at provider list price; the same day local sandboxing went GA in Copilot CLI, the Copilot app and VS Code Agent Host at no extra cost, the CLI learned to discover local Ollama models, and GitHub announced a purpose-built secret-detection model whose new push-protection and /security-review checks will draw AI Credits.** **GPT-6.1 Sol landed in Copilot on OpenAI's launch day (2026-09-29, Pro+/Max/Business/Enterprise), the fifth day-one frontier add in eight days, and on 2026-10-02 the four-model retirement slate executed on schedule: Gemini 3.5 Flash, Gemini 3.6 Flash, Kimi K2.7 Code and Claude Opus 4.7 are gone from every Copilot surface.** Same week: HydraFusion's multi-model orchestration reached VS Code 1.140 and the Copilot app (9/30), **computer use** let Copilot CLI and the Copilot app drive desktop applications on macOS and Windows in public preview (10/01), dynamic workflows defined in code arrived for CLI, app and SDK (10/01), the September VS Code roundup added automations, agent merge and Dev Container sessions (10/01), and Copilot code review gained REST/GraphQL request support with Balanced as the default effort level (10/02). Both have since executed: the six-model slate on 10/19 and the 10/22 default policy for new features for Business and Enterprise. The 9/22 and 9/28 adds (Opus 5.5, GPT-6 Sol/Luna, Sonnet 5.5) all stand, metered at provider list pricing.

Last updated: 2026-11-05Free tier availablePowered by Multi-model picker: GPT-6.1 Terra (Pro and up, added 2026-10-26), GPT-6.1 Sol (Pro+ and up, added 2026-09-29), Claude Sonnet 5.5 (Pro and up, 2026-09-28), Claude Opus 5.5 (Pro+ and up, 2026-09-22), GPT-6 Sol (Pro+ and up, 2026-09-22), GPT-6 Luna (Pro and up, 2026-09-22), Grok 4.7 (all paid SKUs, 2026-09-21), Grok 4.6 (2026-08-14), Gemini 3.7 Flash (2026-08-13, removed 10/19), MAI-Code-1.1-Flash (2026-08-11), Kimi K3 (2026-08-06), Claude Opus 5, GPT-6 Astra, Fable 5.1, Gemini 3.8 Flash and earlier models per the model policy; Copilot Fusion (GA 2026-10-28, successor to HydraFusion and auto model selection) routes and orchestrates them. Gemini 3.5 Flash, Gemini 3.6 Flash, Kimi K2.7 Code and Claude Opus 4.7 were removed 2026-10-02; six more (Gemini 3.7 Flash, GPT-5.5, GPT-5.4, GPT-5.4 mini, GPT-5 mini, Grok 4.5) were removed 2026-10-19

Score Breakdown

9.0
Ease of Use
8.0
Output Quality
8.0
Value
8.0
Features
Visit GitHub Copilot

The Good and the Bad

What we like

  • +Copilot Fusion (GA 2026-10-28, all plans) picks and combines models per task at Standard, High or Max effort, so you stop hand-picking a model for every prompt, and it is 30% off model costs for a limited time
  • +Gemini 4 Argon was in the picker on 2026-10-13, the day Google made it GA, on every Copilot client from VS Code to Xcode and Eclipse -- the fourth frontier model in a row Copilot has carried on launch day
  • +Inline code completions feel magical -- it predicts what you're about to type with scary accuracy
  • +Works directly in VS Code, JetBrains, Neovim -- no context switching
  • +Copilot Chat is great for explaining code, writing tests, and debugging
  • +Free tier is functional enough for personal projects

What could be better

  • −Copilot included with GitHub Enterprise carries no per-seat AI credits: Chat, code review and cloud agents stop when the enterprise budget runs out (completions keep working), so budget alerts now matter more than seat counts
  • −Copilot keeps pruning older frontier models on short notice: Claude Opus 4.8, Claude Opus 5 and GPT-5.6 Terra leave every surface on 2026-11-16, a month after the announcement, and Gemini 4 Argon is not available on the $10 Pro plan
  • −Suggestions can be confidently wrong -- it'll autocomplete plausible-looking code that has subtle bugs
  • −Sometimes suggests code that's clearly copied from public repos, raising licensing questions
  • −Context window for chat is smaller than standalone tools like Claude
  • −Multi-file refactoring understanding is limited compared to Cursor

Pricing

Free

$0
  • ✓2,000 completions/mo
  • ✓50 chat messages/mo
  • ✓VS Code + JetBrains
  • ✓Still open to new signups (2026-04-20)

Pro

$10/month
  • ✓Plans page shows a Get started button with no pause notice as of 2026-09-28 (signups were paused 2026-04-20; no changelog entry has announced a reopening)
  • ✓Claude Opus tier is Pro+ and up (Opus 5.5 added 2026-09-22); Pro gets Claude Sonnet 5.5 (9/28), GPT-6 Luna (9/22) and GPT-6.1 Terra (10/26), plus Copilot Fusion (10/28)
  • ✓Effective 2026-06-01: $10 in monthly AI Credits replaces premium-request quota

Pro+

$39/month
  • ✓Plans page shows a Get started button with no pause notice as of 2026-09-28 (signups were paused 2026-04-20; no changelog entry has announced a reopening)
  • ✓Claude Opus 5.5 (2026-09-22), GPT-6 Sol (9/22), Sonnet 5.5 (9/28), Grok 4.7 (9/21) -- day-one adds
  • ✓Full frontier-model picker
  • ✓Live 2026-06-01: $39 in monthly AI Credits

Copilot Max (NEW 2026-06-01)

$100/month
  • ✓Upgrade for existing Student/Pro/Pro+ subscribers
  • ✓Higher included usage + higher spending limits
  • ✓Listed at $100/month with a Get started button on the plans page as of 2026-09-28
  • ✓Usage-based AI Credits + token metering

Copilot Standalone (NEW 2026-10-28)

$10/user/month
  • ✓For users in an enterprise account who do not hold a GitHub Enterprise licence
  • ✓Same features and models as Copilot Enterprise
  • ✓No monthly AI credits per seat: usage is metered against the enterprise budget
  • ✓Completions and next edit suggestions keep working at zero budget; Chat, code review and cloud agents need budget

Business

$19/user/month
  • ✓Organization management
  • ✓Policy controls
  • ✓Audit logs
  • ✓Effective 2026-06-01: $19 in monthly AI Credits per user

Enterprise

$39/user/month
  • ✓From 2026-10-28, GitHub Enterprise Cloud customers working with a GitHub or Microsoft account team can instead get Copilot included for every licensed user: no per-seat AI credits, usage metered against the enterprise budget from the first request (self-serve later)
  • ✓Org-wide governance + SSO
  • ✓Effective 2026-06-01: $39 in monthly AI Credits per user
  • ✓Standard token metering above the allotment

Known Issues

  • COPILOT USAGE METRICS API: CODE REVIEW 'ADDRESSED RATE' AND TELEMETRY-GAP FLAGS (2026-11-05, vendor-primary): two admin-facing additions for Copilot Business/Enterprise. **(1) Code review outcomes:** repository-level reports gain a `copilot_code_review` object with `total_comments_evaluated` and `total_comments_addressed` -- addressed counts any way the developer acted ('applying the suggestion, asking Copilot to make the fix, editing the code, or replying that the fix will come in a later pull request') -- broken down by severity, review tier (lite, balanced, max), comment type and customization, plus `total_merged_no_human_change_request_after_review`. Comments are counted on the PR's merge day, and bot-opened PRs (such as Copilot cloud agent's) are usually not evaluated. **(2) Telemetry gaps:** per-user `client_telemetry_present` and `usage_metrics_requirements` fields, plus daily counts of users missing client telemetry or running IDE / extension versions below the metrics minimums -- such users 'still count as active, but their reports lack detail such as lines of code, feature breakdowns, and model breakdowns'. **Read for buyers:** this is the first first-party measure of how often Copilot code review comments are acted on, which is the number to watch if you are paying for review on every PR.Source: GitHub Changelog (2026-11-05-usage-metrics-api-adds-code-review-details-and-addressed-rate; 2026-11-05-usage-metrics-api-flags-telemetry-and-version-gaps) -- fetched 2026-11-05 · 2026-11-05
  • COPILOT IN XCODE'S CODING ASSISTANT AND IN JETBRAINS AIR, BYOK IN THE USAGE API, AND THE UNIVERSE MEMORY ITEMS THIS PAGE MISSED (2026-11-03 to 2026-11-04, plus 2026-10-28; vendor-primary): **(1) Xcode (11/04):** 'You can now work with GitHub Copilot directly in Xcode's built-in coding assistant', in public preview 'starting with Xcode 27.3 beta' on macOS Tahoe 26.6 or later; set up under Xcode > Settings > Intelligence; 'All Copilot plans are supported.' The separate Copilot for Xcode extension stays available. **(2) JetBrains Air (11/03):** Copilot appears in Air's agent picker over the Agent Client Protocol; local tasks use your Copilot subscription, and cloud tasks through Air Teams run in an isolated Docker container on a fresh VM and 'use your Copilot plan rather than JetBrains AI credits'. Limits GitHub lists: you cannot choose personal vs organization plan per task, Air's analytics do not include Copilot usage, and Air admins can disable Copilot but cannot set Copilot spending limits. **(3) BYOK reporting (11/03):** the Copilot usage metrics API now separates local BYOK (a developer's own key or local models via Ollama, Foundry Local, LM Studio, llama-server) and enterprise custom models from Copilot-hosted usage; existing totals are unchanged. **(4) From Universe (10/28), not previously recorded here:** 'GitHub Copilot Memory is now generally available on paid Copilot plans', across the Copilot app, CLI, VS Code, cloud agent, code review, Slack, Teams and GitHub Mobile; and **Copilot dreaming** (public preview) reflects on past sessions to curate personal and repository memories and proposes changes to checked-in skills and instructions as pull requests. 'Dream runs consume AI credits.'Source: GitHub Changelog (2026-11-04 Xcode coding assistant; 2026-11-03 JetBrains Air; 2026-11-03 BYOK in the Copilot usage API; 2026-10-28 GitHub intelligence) -- fetched 2026-11-05 · 2026-11-04
  • PLUGIN DEFAULTS AND ENTERPRISE PLUGIN CONTROLS GA (2026-10-28, vendor-primary): 'Repository plugin defaults and enterprise-managed plugin controls are now generally available in VS Code, Copilot CLI, and the GitHub Copilot app.' Maintainers can define a plugin marketplace and default plugins for a trusted repository so contributors skip per-project setup; enterprise admins can supply approved marketplaces and **require or block plugins** through managed settings -- 'Required plugins are automatically installed and enabled, and users can't disable them.' Repository defaults reach individual Copilot users plus Business and Enterprise; managed controls need GitHub Enterprise Cloud with enterprise-issued Business or Enterprise licences. Published on Universe day but after our 10/29 snapshot. GitHub's changelog was otherwise quiet on Copilot from 10/29 to 11/02.Source: GitHub Changelog (github.blog/changelog/2026-10-28-repository-plugin-defaults-and-enterprise-plugin-controls-generally-available, JSON-LD 2026-10-28), fetched 2026-11-02 · 2026-10-28
  • GITHUB UNIVERSE: COPILOT FUSION GA, COPILOT INCLUDED IN GITHUB ENTERPRISE, IMAGES, PROJECTS, CODE REVIEW MAX (2026-10-28, vendor changelog): **(1) Copilot Fusion (GA):** 'Instead of picking a model for each task, choose Copilot Fusion and let Copilot intelligently route and orchestrate which models it uses for you.' It merges Project HydraFusion's multi-model orchestration with auto model selection; effort levels Fusion Standard, High and Max; workflows Single, Cascade (an efficient model drafts, a quality check escalates) and Critique. GitHub says Fusion Max 'scored 60, within two points of Claude Code with Opus 5.5 Max' on the Artificial Analysis Coding Agent Index at about half the API cost per task (vendor-reported). Billing is per model used at standard rates, with 'a 30% discount on model costs' for a limited time on paid plans. Available on all Copilot plans in VS Code, the Copilot app, Copilot CLI and JetBrains; prior auto, HydraFusion and HydraFusion Max selections move to the matching Fusion level. A separate entry adds experimental local-model routing (MAI Code 1.1 Flash via Windows ML; Qwen3.8 Flash Next through a local OpenAI-compatible server in the CLI). **(2) GitHub Enterprise includes Copilot:** enterprises working with a GitHub or Microsoft account team 'can now get Copilot included with GitHub Enterprise for every licensed user'; no monthly AI credits per seat, usage 'metered from the first request against your enterprise budget'; completions and next edit suggestions keep working at zero budget while Chat, code review and Copilot in the cloud need budget. Users without a GitHub Enterprise licence get Copilot Standalone at $10 per user per month. Self-serve enterprises: not yet, dates 'in the coming weeks'. Copilot Business and Enterprise remain on sale. **(3) Images:** generate and edit images in any Copilot chat via a tool 'powered by GPT Image 2.5 Flare', public preview on Pro, Pro+, Business and Enterprise, billed as image tokens against AI credits, optional 'Made with AI' watermark off by default. **(4) Copilot app:** Projects (public preview) run multi-repo and non-code work with as many agents as needed; computer use on macOS and Windows (public preview); a security extension for code scanning alerts (public preview). **(5) Code review:** new Max effort (several review agents across models from different providers plus an aggregation agent; public preview, uses more AI credits) and Copilot approvals GA (an approval counts toward required approvals and is dismissed on new commits). **(6) Agent Merge** now fixes failing Actions workflows and merge conflicts from the pull request merge box on github.com, for paid plans. **(7) One cloud experience (GA on paid plans):** a session started on web, GitHub Mobile, CLI, the Copilot app, VS Code, Slack or Teams continues on any of them; cloud sandbox compute can draw on the included Actions entitlement. **(8) Connectors:** 26 GitHub-verified MCP services (Slack, Figma, Notion, Sentry, Linear, Atlassian Rovo, Cloudflare, Vercel, Work IQ and others) in public preview, connect once for all Copilot surfaces. **(9) Copilot Automations:** API triggers, 24 new GitHub event triggers, chaining, 20 new tools plus custom MCP tools, and AI credit limits per run, day, week or month; shared team automations rolling out on Pro, Pro+, Max, Business and Enterprise. **(10) GitHub Copilot SDK is becoming Microsoft Copilot SDK** (phased rename, no action needed). Also: GitHub Agentic Workflows GA, and Replica, an open-source AI prototyping tool from GitHub Design.Source: GitHub Changelog 2026-10-28: route-and-orchestrate-models-with-github-copilot-fusion, github-copilot-fusion-now-supports-selection-of-local-models, github-enterprise-now-includes-access-to-github-copilot, generate-and-edit-images-directly-in-copilot, github-copilot-app-universe-releases, copilot-code-review-merge-with-confidence-with-max-and-approvals, agent-merge-watches-and-fixes-pull-requests-in-the-cloud, work-from-anywhere-with-a-new-connected-copilot-cloud-experience, connect-your-tools-to-copilot-without-all-the-setup, new-tools-triggers-and-cost-controls-for-copilot-automations, github-copilot-sdk-is-becoming-microsoft-copilot-sdk, github-agentic-workflows-generally-available, introducing-replica-an-open-source-ai-prototyping-tool (JSON-LD datePublished 2026-10-28) · 2026-10-28
  • GPT-6.1 TERRA ON LAUNCH DAY, AND THE OCTOBER DATED CHANGES EXECUTED (2026-10-19 to 2026-10-27, vendor changelog): **GPT-6.1 Terra (10/26):** 'now generally available in GitHub Copilot ... designed for everyday agentic coding and terminal work at a lower cost than GPT-6.1 Sol', billed at provider list price, for Copilot Pro, Pro+, Max, Business and Enterprise (it is on the $10 Pro plan, unlike Gemini 4 Argon) in VS Code, Visual Studio, Copilot CLI, cloud agent, the Copilot app, github.com, GitHub Mobile, JetBrains, Xcode and Eclipse; gradual rollout. **Six-model removal (10/19, executed):** 'As of today, October 19, 2026, we have deprecated the following models across most GitHub Copilot experiences': Gemini 3.7 Flash, GPT-5.5, GPT-5.4, GPT-5.4 mini, GPT-5 mini and Grok 4.5. The suggested alternatives changed from the 9/18 notice: GPT-5.5 and GPT-5.4 now point to GPT-6.1 Sol (was GPT-5.6 Sol), GPT-5.4 mini and GPT-5 mini to GPT-6 Luna (was GPT-5.6 Luna), Grok 4.5 to Grok 4.7 (was Grok 4.6); Gemini 3.7 Flash still to Gemini 3.8 Flash. **Default policy for new features (10/22):** 'we're gradually rolling out enforcement of the policy through October 26'; unconfigured eligible GA features and MCP-server settings follow the chosen global default (Enabled, Disabled or Let organizations decide) as a one-time change, explicit choices are kept and previews stay opt-in. **Also:** Copilot model policy REST API GA at enterprise, enterprise-team, agent-app and organization level (10/19); Industry comparison benchmarks in the Copilot impact dashboard, opt-in (10/22); a consolidated GitHub app for Microsoft Teams with Copilot, and the legacy GitHub app (to be renamed 'GitHub (deprecated)') retires November 16 (10/23); assisted permissions GA in Copilot CLI, the Copilot app and VS Code (10/26); server-to-server authentication for Copilot SDK apps on organizations, public preview, billed to the organization (10/27).Source: GitHub Changelog: 2026-10-26-gpt-6-1-terra-in-github-copilot, 2026-10-19-github-copilot-models-deprecated-on-october-19-2026 (datePublished 2026-10-19T14:27 PDT), 2026-10-22-global-default-for-copilot-features-generally-available, 2026-10-19-copilot-model-policy-api-for-enterprise-orgs-and-teams, 2026-10-22-industry-comparisons-in-the-copilot-impact-dashboard, 2026-10-23-turn-notifications-into-code-with-github-copilot-in-teams, 2026-10-26-github-copilot-assisted-permissions-generally-available, 2026-10-27-github-copilot-supports-server-to-server-authentication · 2026-10-26
  • SLACK + TEAMS (2026-10-15): **Multi-repository support for GitHub in Slack and Teams** -- @GitHub can create issues, make code changes and open PRs, track your open PRs and investigate bugs across several repositories in one conversation without switching. Also: mention @GitHub anywhere in a message (not just at the start), smoother repository switching, shared Slack messages and threads usable as context, and reliability fixes for task startup and interruption recovery.Source: github.blog/changelog/2026-10-15-copilot-in-slack-and-teams-adds-model-and-repository-updates · 2026-10-17
  • GEMINI 4 ARGON DAY ONE, THREE MODELS REMOVED ON NOVEMBER 16, HYDRAFUSION MAX, AND TWO CODE-SECURITY AI FEATURES GO GA (2026-10-13 to 2026-10-15, vendor changelog): **(1) Gemini 4 Argon (10/13):** 'now generally available in GitHub Copilot ... designed for complex coding, agentic, and reasoning tasks'; GitHub says it 'solved more tasks than previous Gemini models on every coding benchmark, with its largest gains in Copilot CLI'. 'Billed at provider list pricing under usage-based billing' (Google lists $2 / $10 per 1M until 11/30, then $4 / $20). Available to **Copilot Pro+, Max, Business and Enterprise -- not Pro** -- in VS Code, Visual Studio, Copilot CLI, the coding agent, the Copilot app, JetBrains, Xcode and Eclipse; gradual rollout; on by default under default model enablement. That is the fourth frontier launch in a row Copilot has carried on day zero or one. **(2) Removals announced 10/15, effective 2026-11-16, across Chat, inline edits, ask and agent modes and completions:** Claude Opus 4.8 -> Claude Opus 5.5 ('will remain available to individual GitHub Copilot subscribers on annual plans so they retain an Opus offering'); Claude Opus 5 -> Claude Opus 5.5; GPT-5.6 Terra -> GPT-6.1 Sol or GPT-6 Luna. Alternatives are auto-enabled for Business/Enterprise unless an admin turned off the global default. **(3) HydraFusion update + HydraFusion Max (10/14, research preview):** the multi-model orchestrator's pool now includes Claude Opus 5.5, GPT-6 Astra, GPT-6.1 Sol and GPT-6 Luna; it now tailors effort per stage (medium for drafts, high for review); Cascade adds a judging step and Critique uses parallel critics that inspect the workspace with tools. Max applies higher effort for 'the deepest possible reasoning'. Both run in VS Code, the Copilot app and CLI on Pro, Pro+, Business and Enterprise with preview features enabled. **(4) Agentic autofix for code scanning alerts GA (10/15):** assign a CodeQL alert (or up to 25 at once) to Copilot, which explores the codebase, reruns CodeQL to confirm the fix and opens a draft PR; Code Security / Advanced Security customers on GitHub Enterprise Cloud; replaces the free Generate Fix button when enabled; consumes Code Security AI Credits plus Actions minutes and needs a Copilot billing method. **(5) AI Scan for pull requests GA (10/15):** AI findings alongside CodeQL for languages CodeQL does not cover (PHP, Shell, Terraform HCL, Dockerfiles, JSP, Blazor); advisory only, cannot block merges; free on public repos, Code Security or Advanced Security required for private; **no Copilot license required**; consumes Code Security AI Credits. **(6) 10/13:** secret scanning now blocks new Anthropic key types with push protection by default.Source: GitHub Changelog: 2026-10-13-gemini-4-argon-in-github-copilot, 2026-10-15-upcoming-deprecation-of-selected-github-copilot-models-in-mid-november, 2026-10-14-hydrafusion-updates-and-new-max-option, 2026-10-15-agentic-autofix-for-code-scanning-alerts-general-availability, 2026-10-15-ai-scan-for-pull-requests-generally-available, 2026-10-13-secret-scanning-blocks-anthropic-keys-by-default-adds-salesforce (JSON-LD datePublished on each) -- fetched 2026-10-15 · 2026-10-15
  • OCTOBER 6-12: HAIKU 5.5 DAY ONE, LOCAL SANDBOXING GA, LOCAL MODELS IN THE CLI, A SECRET-DETECTION MODEL, CODE REVIEW BILLING CONTROLS, A JETBRAINS FLOOR RAISE AND A METRICS UNDERCOUNT (vendor-primary, GitHub changelog): **(1) Claude Haiku 5.5 GA (10/07, launch day)** -- 'designed for fast, high-volume work like subagents, quick edits, and terminal tasks'; GitHub says it 'matched Claude Sonnet 5 on many coding tasks while using significantly fewer tokens and steps'. **Billed at provider list pricing under usage-based billing**; Copilot Pro, Pro+, Max, Business and Enterprise; VS Code, Visual Studio, Copilot CLI, cloud agent, the Copilot app, github.com, GitHub Mobile, JetBrains, Xcode and Eclipse; gradual rollout; on by default for Business/Enterprise unless the admin turned off default enablement. **(2) Local sandboxing GA (10/07)** in Copilot CLI, the Copilot app and VS Code sessions using Agent Host: agent-run commands get restricted filesystem, network, Git and GitHub CLI credential access by developer or organization policy, powered by **Microsoft eXecution Container (MXC)** across Windows, macOS and Linux; enterprises can require it via managed settings developers cannot weaken; **included at no additional cost**. **(3) Local model discovery in Copilot CLI (10/07, CLI 1.0.94-0+)** -- `/model` lists models from a running local Ollama instance; models must support tool calling and streaming; choosing one does NOT turn on offline mode (`COPILOT_OFFLINE=true` is separate) or disable telemetry. Intelligent routing with local models was announced but is not yet available. **(4) Purpose-built secret-detection model (10/07)** -- GitHub's fine-tuned model now powers AI-detected password alerts (automatic upgrade, still included in GHSP/GHAS); **AI push protection is in private preview and AI secret scanning in the Copilot /security-review command is 'available soon in private preview' -- both will consume GitHub AI Credits, with credit usage 'introduced in the coming weeks'**; AI-detected alerts come to GHES 3.23 in public preview. **(5) Copilot code review billing (10/08)** -- org owners can bill reviews requested by licensed members to the organization's cost center instead of the member's quota (requires AI Credits paid usage, optional budget; default is still the member's entitlement, and a review fails when that quota is exhausted), and can require that review requests come from a Copilot license provided by the org or enterprise (blocks personal licenses). **(6) Weekly roundup (10/09)** -- the Copilot app can use one GitHub account for the Copilot license and another for repositories; VS Code 1.141 adds a grid Agents window and worktree cleanup. **(7) JetBrains (10/10)** -- enterprise-managed default models, a Fix action from diagnostics, an MCP auto-start toggle for Copilot and Claude, and **support for JetBrains IDE 2025.1 ended -- 2025.2 or later is now required**. **(8) Usage metrics undercount (10/06)** -- IDEs that moved agent sessions to the Copilot SDK stopped identifying themselves, so agent activity and agent lines of code were undercounted (some counted as Copilot CLI); VS Code 1.139.0+ fixes it, Visual Studio 18.12 (October), JetBrains (late October), Eclipse and Xcode (by November) to follow; **billing was not affected and missing data cannot be backfilled**. **(9) Telemetry identity (10/12)** -- enterprise admins can attribute OpenTelemetry Copilot activity to a user (login, OS account, host name) without capturing prompts, and separately enable five content classes; all off by default. **Calendar note:** GitHub's 9/03 announcement that existing card/PayPal Business and Enterprise customers move to upfront per-seat charging from 2026-10-01 has had no follow-up changelog entry through 10/12; this page treats it as announced and not independently confirmed as executed.Source: GitHub Changelog entries fetched 2026-10-12: 2026-10-07-claude-haiku-5-5-in-github-copilot, 2026-10-07-local-sandboxing-for-github-copilot-now-generally-available, 2026-10-07-discover-local-models-in-github-copilot-cli, 2026-10-07-purpose-built-model-for-leaked-secret-detection, 2026-10-08-copilot-code-review-new-organization-billing-options-and-controls, 2026-10-09-github-copilot-weekly-releases-october-5, 2026-10-10-new-controls-and-chat-improvements-in-copilot-for-jetbrains, 2026-10-06-update-your-ide-to-restore-agent-activity-in-copilot-usage-metrics, 2026-10-12-client-telemetry-user-attribution-and-configurable-capture · 2026-10-07
  • DATED FLIP EXECUTED -- THE 2026-10-02 SLATE REMOVED FOUR MODELS ON SCHEDULE, AND CODE REVIEW GOT AN API THE SAME DAY (2026-10-02, vendor-primary): GitHub's 'Retired' entry: 'As of today, October 2, 2026, we have deprecated the following models across all GitHub Copilot experiences (including Copilot Chat, inline edits, ask and agent modes, and code completions)': **Gemini 3.5 Flash -> Gemini 3.8 Flash; Gemini 3.6 Flash -> Gemini 3.8 Flash; Kimi K2.7 Code -> Kimi K3; Claude Opus 4.7 -> Claude Opus 5.5.** Note the Opus replacement changed between notice and execution: the 9/03 notice named Claude Opus 5, the 10/02 execution names **Opus 5.5** (launched 9/22). 'No action is required to remove the deprecated models'; Enterprise admins may need to enable the alternatives through model policies. That is four of the ten models out; the remaining six (Gemini 3.7 Flash, GPT-5.5, GPT-5.4, GPT-5.4 mini, GPT-5 mini, Grok 4.5) went on 10/19. **Copilot code review (same day, 'Improvement'):** 'You can now request a GitHub Copilot code review through the REST and GraphQL APIs and set the review effort level for each request' -- GA for Pro, Pro+, Max, Business and Enterprise -- and **'Balanced is now the default review effort level'** for new and existing repositories and organisations (an explicit Lite selection is respected); GitHub says this default change 'took effect September 28, 2026' as announced on 8/28. Effort can be set at enterprise (AI controls > Agents > Copilot code review), organisation, repository and personal level, each overriding the one above.Source: GitHub Changelog (github.blog/changelog/2026-10-02-selected-models-in-github-copilot-deprecated/ and github.blog/changelog/2026-10-02-copilot-code-review-api-support-and-new-default-effort-level/) -- fetched 2026-10-02 via curl with browser UA · 2026-10-02
  • GPT-6.1 SOL ON DAY ONE, HYDRAFUSION IN VS CODE AND THE APP, COMPUTER USE FOR DESKTOP APPS, DYNAMIC WORKFLOWS, AND THE SEPTEMBER VS CODE ROUNDUP (2026-09-29 to 2026-10-01, all vendor-primary): **(1) GPT-6.1 Sol (9/29)** -- 'the latest model from OpenAI, is now generally available and rolling out in GitHub Copilot' for agentic coding and terminal workflows; GitHub's early testing: 'it reliably completed tasks while using noticeably fewer tokens and steps than earlier models in the GPT-6 and GPT-5.6 families'. **Billed at provider list pricing under usage-based billing; Pro+, Max, Business and Enterprise**; in the picker across VS Code, Visual Studio, Copilot CLI, the coding agent, the Copilot app, github.com, Mobile, JetBrains, Xcode and Eclipse; gradual rollout. OpenAI's own rate card for the model is $2 / $0.10 cached / $10 per 1M. That makes five frontier models on their vendors' launch day in eight days (Opus 5.5, GPT-6 Sol, GPT-6 Luna on 9/22; Sonnet 5.5 on 9/28; GPT-6.1 Sol on 9/29). **(2) HydraFusion research preview in VS Code 1.140+ and the Copilot app (9/30)** -- 'expanding beyond Copilot CLI'. It sits in the model picker but 'rather than being a single model, it orchestrates multiple models', choosing one of three workflows: **Single** (one model solves it), **Cascade** (an efficient model drafts and a quality gate escalates to a stronger model if needed) or **Critique** (one model drafts, a read-only critic from a different model family reviews, the drafter revises once). Enable with chat.copilot.hydraFusion.enabled in VS Code or Settings > HydraFusion in the app; org and enterprise admins may need to allow preview features. Also more step transparency and real-time progress. **(3) Computer use (10/01, public preview)** -- Copilot CLI and the Copilot app on macOS and Windows can 'interact with desktop applications on your behalf (e.g., reading accessible app content and visual context, clicking controls, entering and editing text, pressing keys, scrolling, dragging, and navigating workflows across applications)', aimed at legacy and GUI-only software with no API, CLI or MCP. Copilot asks approval before controlling an app, always-allow lists are reviewable, macOS walks through Accessibility and Screen Recording permissions, and org-managed settings can disable it. Toggle with /computer on | show | off. **(4) Dynamic workflows (10/01)** -- for Copilot CLI, the Copilot app and the Copilot SDK: 'a program that defines how a task is carried out', combining automated steps with one or more agents in sequence or parallel, with structured hand-offs, subagents verifying each other, user prompts and pause-and-resume checkpoints; the program lives inside a Copilot extension. Distinct from /fleet (ad-hoc delegation); Copilot can author workflows for you. **(5) VS Code September 2026 roundup (10/01, v1.136-v1.140):** HydraFusion in the model picker (preview), scheduled automations (preview), **agent merge** that handles review feedback, failed checks, merge conflicts and workflow reruns (preview), PR creation from Copilot, Claude or Codex agent sessions, hierarchical related chats and sessions, Mark as Done and auto-cleanup, an application badge for sessions needing attention, Dev Container sessions from the Agents window incl. SSH/Tunnel/WSL, workspace-less quick chats, **'pick up a Codex conversation from the ChatGPT app in VS Code'**, uninterrupted active chats and GitHub issue/PR context attachment.Source: GitHub Changelog (github.blog/changelog/2026-09-29-gpt-6-1-sol-in-github-copilot/, .../2026-09-30-hydrafusion-in-vs-code-and-the-github-copilot-app/, .../2026-10-01-github-copilot-can-now-interact-with-desktop-apps/, .../2026-10-01-dynamic-workflows-in-copilot-cli-and-the-copilot-app/, .../2026-10-01-github-copilot-in-vs-code-september-2026-releases/) + developers.openai.com/api/docs/pricing.md (gpt-6.1-sol) -- all fetched 2026-10-02 via curl with browser UA · 2026-10-01
  • CLAUDE SONNET 5.5 IN COPILOT ON ANTHROPIC'S LAUNCH DAY -- PRO AND UP, GA, AUTO-ENABLED (2026-09-28, vendor-primary): GitHub added **Claude Sonnet 5.5** as 'generally available in GitHub Copilot', 'designed for well-scoped everyday work like building features and fixing bugs'; in GitHub's early testing it matched Claude Sonnet 5 on coding tasks 'while using significantly fewer steps, tokens, and tool calls' and 'finished tasks noticeably faster'. **Availability: Copilot Pro, Pro+, Max, Business and Enterprise** (one tier wider than Opus 5.5, which excludes Pro) across VS Code, Visual Studio, Copilot CLI, the coding agent, the Copilot app, github.com, GitHub Mobile, JetBrains, Xcode and Eclipse; gradual rollout. Billed at provider list pricing under usage-based billing (Anthropic's list: $2/$10 per 1M). Under default model enablement it is on automatically unless an administrator turned off the global default or disables it. Fourth day-one model add in a week (Opus 5.5, GPT-6 Sol, GPT-6 Luna on 9/22).Source: GitHub Changelog (github.blog/changelog/2026-09-28-claude-sonnet-5-5-in-github-copilot, datePublished 2026-09-28) + anthropic.com/claude-sonnet-5-5 (9/28) -- fetched 2026-09-28 via curl with browser UA · 2026-09-28
  • OPUS 5.5, GPT-6 SOL AND GPT-6 LUNA ALL LAND IN COPILOT ON 2026-09-22 -- THE DAY BOTH VENDORS SHIPPED THEM (vendor-primary): two changelog entries the same day. **Claude Opus 5.5** -- 'for agentic coding, long-running agentic tasks, and knowledge work'; in early testing it 'resolved tasks comparably to Claude Opus 5 while using significantly fewer steps and tokens' and 'quickly recovered from errors in multistep tasks'; **Opus 5.5 watermarks its text outputs** (no change to meaning, tokens or cost -- GitHub links Anthropic's text-watermark explainer); **Pro+, Max, Business and Enterprise** (not Pro). **GPT-6 Sol** -- 'a balanced model for interactive and agentic coding ... careful, multistep validation' -- **Pro+, Max, Business and Enterprise**; **GPT-6 Luna** -- 'lightweight, cost-efficient ... the lowest-cost option in the GPT-6 family' -- **Pro, Pro+, Max, Business and Enterprise**. All three: usage-based billing at provider list prices (Anthropic $4/$20; OpenAI $2/$10 and $0.10/$0.50), every surface from VS Code to Eclipse, gradual rollout, auto-enabled under default model enablement unless an admin disabled the global default. The 9/25 weekly roundup confirms the same tier map and adds Grok 4.7 for all five paid SKUs. Anthropic's Opus 5.5 post quotes GitHub CPO Mario Rodriguez: 'in our testing across GitHub Copilot CLI and VS Code, Claude Opus 5.5 used among the fewest tokens and steps we measured. In VS Code, it solved more terminal tasks than Opus 5 in less than half the steps.' OpenAI's GPT-6 caching post quotes him again: prompt-caching improvements cut the share of prompt tokens needing fresh processing 'by more than 50%' across billions of Copilot requests. Picker context: the 10/02 and 10/19 slates removed ten models; the additions this week are the replacements users will actually land on.Source: GitHub Changelog (2026-09-22-claude-opus-5-5-is-now-available-in-github-copilot; 2026-09-22-openais-gpt-6-sol-and-gpt-6-luna-now-available; 2026-09-25-github-copilot-weekly-releases-september-21) + anthropic.com/claude-opus-5-5 + openai.com/index/better-prompt-caching-for-gpt-6/ -- fetched 2026-09-28 via curl with browser UA · 2026-09-22
  • NEW DATED FLIP -- 2026-10-22: A GLOBAL 'DEFAULT POLICY FOR NEW FEATURES' TAKES EFFECT FOR COPILOT BUSINESS AND ENTERPRISE (announced 2026-09-24, vendor-primary): GitHub introduced 'a new global default policy for generally available GitHub Copilot features and supported client capabilities' in enterprise and organization settings. 'For the next 28 days, you can configure this policy, but it won't affect feature access for your users yet.' Where: AI Controls page > Copilot subpage > 'Default policy for new features'. Scope: eligible features on the enterprise 'Features & clients' page, the Copilot Code Review policy on the Agents page, and the MCP servers in Copilot policy. Options: **Enabled** (current and future eligible features available by default), **Disabled** (current eligible features stay unavailable; future ones need admin approval), **Let organizations decide**. **'Starting October 22, the policy takes effect'**: eligible GA features left Unconfigured follow the global default; explicit enable/disable decisions are preserved; preview features remain opt-in and an existing preview choice survives GA. This is the feature-side twin of the 8/26 model policy that already auto-enables new models. It resolves the older 'no earlier than 9/28 unified Copilot default-on' watch item as a 10/22 effective date with a 28-day configuration window.Source: GitHub Changelog (github.blog/changelog/2026-09-24-default-enablement-of-copilot-features-for-copilot-business-and-enterprise, datePublished 2026-09-24) -- fetched 2026-09-28 via curl with browser UA · 2026-09-24
  • AGENT-PLATFORM WEEK: LOCAL SANDBOXING, OPENTELEMETRY, CODE-REVIEW SETTINGS ON EVERY PLAN, JETBRAINS 1.18, C++ INDEXING, SLACK/TEAMS, AUTOFIX MEMORY, REVIEW-STAGE METRICS (2026-09-22 to 2026-09-25, vendor-primary): **Local sandboxing in the Copilot app (9/23, public preview)** -- per-project filesystem (extra read/write, read-only and denied folders), network (outbound internet, local network) and credential (git HTTPS, GitHub CLI) policy for local repository and working-tree sessions; off by default ('Sandbox new sessions' or `/sandbox on` for a live session); enterprise-managed settings can tighten it; if the OS cannot enforce the policy the sandboxed shell fails rather than running unsandboxed; cloud and remote-host sessions excluded; app and CLI sandboxes configured separately. **OpenTelemetry in the Copilot app (9/22)** -- configure `telemetry` in enterprise `managed-settings.json` to export agent activity (model requests, tool use, step traces) to compatible monitoring; prompt and response content excluded by default. **Copilot code review (9/23, GA)** -- a dedicated personal 'code review' settings page on **every** plan including Business and Enterprise (previously Pro/Pro+/Max only): automatic review on PR create / coauthor / undraft, on new pushes, on drafts; a default review effort (Lite or Balanced); plus an enterprise-wide default review effort inherited by organization-owned repositories. **Copilot for JetBrains 1.18.0 (9/22)** -- assisted approvals (public preview: low-risk tool calls auto-approved, higher-risk prompt you), re-edit an earlier message and Copilot rewinds conversation and file changes, organization/enterprise skills and managed instructions in local and agent sessions, plan mode for the Codex agent, per-tool MCP controls and a toggle for the built-in GitHub MCP server, side-by-side chat panel. **Copilot CLI C++ (9/22)** -- whole codebase indexing (WCI) with the Microsoft C++ Language Server, on by default, `/lsp logs` for progress. **Slack and Teams (9/25, public preview for Business/Enterprise)** -- switch models mid-conversation and keep the choice for the thread (Slack channel defaults, default owners/repos), Slack files/attachments/message links and Teams inline images/forwarded messages as context, duplicate-issue check before creating issues, links back to the source discussion. **Agentic autofix uses Copilot Memory (9/25, both public preview)** -- reads memories to resolve security alerts and stores fix patterns that other Copilot features (code review, cloud agent) can learn from. **Usage metrics API (9/25)** -- `pull_request_review_times` on repos-1-day rows: median and p90 minutes from ready-for-review to first review, first to final review, and final review to merge; human reviews only, no backfill. **Weekly roundup (9/25)** also lists VS Code 1.139 running agents in Dev Containers on SSH, Tunnel and WSL hosts. **Plans page check (2026-09-28):** github.com/features/copilot/plans lists Free $0, Pro $10, Pro+ $39 and **Max $100 per month**, each with a 'Get started' button and no pause or waitlist wording -- the first time since the 2026-04-20 pause that we have seen individual tiers presented as purchasable; GitHub has published no changelog entry announcing the reopening, so this is a page observation, not an announcement.Source: GitHub Changelog (2026-09-23-local-sandboxing-in-the-github-copilot-app; 2026-09-22-opentelemetry-in-the-github-copilot-app; 2026-09-23-copilot-code-review-more-ways-to-request-and-configure-reviews; 2026-09-22-new-features-and-improvements-in-copilot-for-jetbrains; 2026-09-22-faster-c-code-intelligence-with-whole-codebase-indexing; 2026-09-25-updates-to-github-copilot-for-slack-and-microsoft-teams; 2026-09-25-agentic-autofix-now-uses-copilot-memory; 2026-09-25-usage-metrics-api-adds-pull-request-review-stages; 2026-09-25-github-copilot-weekly-releases-september-21) + github.com/features/copilot/plans -- all fetched 2026-09-28 via curl with browser UA · 2026-09-25
  • GROK 4.7 IN COPILOT ON XAI'S LAUNCH DAY -- ZERO-DAY PLACEMENT, ON BY DEFAULT (2026-09-21, vendor-primary): GitHub added **Grok 4.7** for **Copilot Pro, Pro+, Max, Business and Enterprise** across VS Code, Visual Studio, Copilot CLI, the Copilot cloud agent, the Copilot app, JetBrains, Xcode and Eclipse, describing it only as 'xAI's latest reasoning model' for 'agentic coding and complex, multistep workflows'. It bills 'at provider list pricing under usage-based billing' -- xAI's list is $2/$6 per 1M, unchanged from Grok 4.6 -- and rollout is 'gradual'. **The posture is the story:** for Grok 4.6 on 8/14 GitHub wrote that the org policy was off by default; for 4.7 the entry says new models are **enabled automatically unless administrators have disabled them**, which is the 8/26 global model policy doing its job. The picker now carries Fable 5.1 (still off by default for data-retention reasons), GPT-6 Astra, Gemini 3.8 Flash, Grok 4.7 and Grok 4.6 -- while Grok 4.5 was removed on 10/19 (next entry). No premium-request multiplier and no benchmark claim was published.Source: GitHub Changelog (github.blog/changelog/2026-09-21-grok-4-7-is-now-available-in-github-copilot, on-page 'September 21, 2026') -- fetched 2026-09-21; xAI list price from docs.x.ai/docs/pricing the same day · 2026-09-21
  • SECOND DEPRECATION SLATE IN SIXTEEN DAYS (EXECUTED 2026-10-19, WITH UPDATED ALTERNATIVES; SEE THE 2026-10-26 ENTRY) -- SIX MORE MODELS OUT ON 2026-10-19, INCLUDING GPT-5.5 LESS THAN SIX MONTHS AFTER IT LAUNCHED (2026-09-18, vendor-primary): GitHub will remove, 'across all GitHub Copilot experiences (including Copilot Chat, inline edits, ask and agent modes, and code completions)' on **October 19, 2026**: **Gemini 3.7 Flash -> Gemini 3.8 Flash; GPT-5.5 -> GPT-5.6 Sol; GPT-5.4 -> GPT-5.6 Sol; GPT-5.4 mini -> GPT-5.6 Luna; GPT-5 mini -> GPT-5.6 Luna; Grok 4.5 -> Grok 4.6.** 'No action is required to remove the models'; for Business/Enterprise the alternatives 'are automatically enabled by default unless administrators have disabled global defaults', otherwise admins enable them under model policies. **Read with the 10/02 slate** (Gemini 3.5/3.6 Flash, Kimi K2.7 Code, Claude Opus 4.7) this is **ten models leaving the picker in seventeen days**, and two of them -- Gemini 3.7 Flash (added 8/13) and GPT-5.5 (GA 4/23) -- were current-generation models this summer. The practical lesson for teams that pin models in enterprise-managed defaults: a pinned default that is on a slate silently breaks on the date. **ALSO SHIPPED 9/17-9/18:** (1) **Copilot code review, improved review experience (9/18, GA)** -- the overview comment now groups findings into Open, Resolved since last review and Previously missed; each comment carries a title; auto-resolution now records a reason ('Won't Fix' or 'Incorrect') based on your later commits and respects replies meant to keep an issue open; accepting a batch of suggestions generates a commit title and description. (2) **Weekly releases for the week of 9/14 (published 9/18):** auto-model-selection tiers reach VS Code, the CLI and the app; VS Code Agents usage analytics for Business/Enterprise; a Sentry canvas in the Copilot app; **VS Code 1.138** adds local Dev Containers for agents (gradual, needs Docker), automatic session cleanup on PR merge and PR creation from Agent Host sessions. (3) **Impact dashboard feature engagement (9/17)** -- enterprise and org 28-day aggregate reports now count active users who used each feature (completions, agent edit, code review, cloud agent, CLI, app) on at least two days. (4) **Usage metrics API (9/17)** now reports agentic CLI customisations -- top-five skills, custom agents, MCP servers, slash commands and plugins with interaction counts, plus distinct-use counts.Source: GitHub Changelog (github.blog/changelog/2026-09-18-upcoming-deprecation-of-selected-github-copilot-models-in-mid-october; 2026-09-18-copilot-code-review-an-improved-review-experience; 2026-09-18-github-copilot-weekly-releases-september-14; 2026-09-17-copilot-impact-dashboard-now-shows-feature-engagement; 2026-09-17-agentic-cli-customizations-now-in-the-usage-metrics-api) -- all fetched 2026-09-21 · 2026-09-18
  • COPILOT BUDGET INCREASE REQUESTS ARE GA -- MEMBERS WHO HIT THEIR AI-CREDIT LIMIT CAN NOW ASK FOR MORE INSTEAD OF BEING BLOCKED (2026-09-16, vendor-primary): previously 'when a member used all the Copilot AI credits available to them, they were blocked from Copilot features that consume credits'. Now a request flow appears at the moment of the limit; org owners, enterprise owners and billing managers see pending requests under 'Requests from members', set a new amount and click **Approve and increase**, which 'immediately restores their access to AI credits'. Requests route automatically to whichever account pays for the budget (org- or enterprise-owned). **Scope: Copilot Business and Copilot Enterprise under usage-based billing.** Relevant to the 10/01 move to upfront per-seat charging already on this page: the credit ceiling remains a hard stop for the individual until an admin acts.Source: GitHub Changelog (github.blog/changelog/2026-09-16-copilot-budget-increase-requests-are-generally-available/, on-page 'September 16, 2026') -- fetched 2026-09-17 · 2026-09-16
  • COPILOT SUGGESTS ALLOWED VALUES FOR REPOSITORY CUSTOM PROPERTIES -- PUBLIC PREVIEW, WITH ITS OWN ADMIN POLICY (2026-09-15, vendor-primary): when an enterprise or organisation admin defines a custom property (the governance metadata used to scope rulesets), Copilot now proposes allowed values -- a multi-select 'FedRAMP' property surfaces compliance values, a single-select 'internet-facing' surfaces yes/no -- accepted in one click. **Business and Enterprise plans; controlled by a new 'Repository custom property suggestions' Copilot policy.** Minor as a feature, notable as the pattern: Copilot is being wired into platform-governance surfaces, each with a separate enablement policy admins must track.Source: GitHub Changelog (github.blog/changelog/2026-09-15-github-copilot-suggests-custom-properties-definitions/, on-page 'September 15, 2026') -- fetched 2026-09-17 · 2026-09-15
  • MAI-CODE-1-FLASH RETIRED ON SCHEDULE, AUTO MODEL SELECTION GETS COST/QUALITY TIERS, AND ENTERPRISE GUARDRAILS ARRIVE FOR AGENT OPERATIONS -- SEVEN COPILOT CHANGELOG ENTRIES 2026-09-08 THROUGH 2026-09-14 (all vendor-primary): **(1) DATED FLIP EXECUTED -- MAI-Code-1-Flash deprecated 2026-09-10** 'across all GitHub Copilot experiences (including Copilot Chat, inline edits, ask and agent modes, and code completions)', suggested alternative **MAI-Code-1.1-Flash**; Enterprise admins may need to enable the alternative via model policies. This closes the one-month forced migration flagged on this page since 8/11 and is the second model in the 9/01-10/02 retirement run. **(2) Auto model selection tiers (9/14):** auto now offers **efficiency, balance and intelligence** tiers that set how cost, quality and latency are weighed per prompt; all three draw from the same model set and 'usage is charged based on the model auto selects, regardless of tier', with the **10% discount for paid subscribers on auto-billed usage retained**. Rolling out in VS Code, Copilot CLI and the Copilot app. GitHub calls it 'the first step' toward user-customisable selection. **(3) Enterprise managed permissions for agent operations (9/09, GA):** Copilot Business/Enterprise admins can now centrally set which agent operations are **blocked, require human approval, or proceed without a prompt**, covering **shell commands, file reads/edits and network domains**; managed restrictions 'can't be weakened by user or workspace settings, auto-approval, or previously saved approvals', with per-team policies. Applies to the Copilot app, CLI and VS Code sessions using Agent Host. **(4) Enterprise-managed sandbox in Copilot for JetBrains (9/08, public preview):** managed policies for sandbox enablement, filesystem/network access, proxy, developer-tool and macOS Keychain access; same release adds cross-file cursor jumps in next-edit suggestions, global project context in chat, enterprise policy diagnostics, subagent model selection, and **`/ide` in Copilot CLI to bind a terminal session to the JetBrains IDE context** (public preview). **(5) Copilot code review (9/11):** it now **auto-resolves its own comments** once a later commit addresses them, writes smart commit messages when you apply a suggestion, runs **the full Copilot SDK shell toolset behind the agent firewall** to validate code under review (builds, tests, scripts), and the **Lite effort level now uses an ensemble of agents** -- GitHub's own experiment numbers: addressed comments up 47% (high severity), 31% (medium), 11% (low) at ~8% lower review cost. **(6) Weekly release, Sept 7 (9/10):** **Jira integration in the Copilot app** (bring issues into a shared canvas and carry context through investigation, implementation and PR prep); **Project HydraFusion in Copilot CLI `/experimental`** -- 'automated semantic routing between local, cloud, and compound models', selected like any other model; **VS Code 1.137**: scheduled recurring agent tasks (hourly/daily/weekly or on demand, public preview), an experimental **voice mode** to talk to and redirect Copilot mid-task, and issue/PR review inside the Agents window without the repo open. **(7) Usage metrics (9/11):** the Copilot usage metrics API now reports dedicated VS Code Agents-window activity (`daily_active_vscode_agent_users`, `totals_by_vscode_agent`) separately from editor Agent Mode. **Also 9/09, adjacent:** bulk **agentic autofix** for GitHub Code Quality findings -- assign up to 25 findings to Copilot at once, it fixes on a branch and opens a PR; consumes AI credits (Team and Enterprise Cloud). **NET READ:** the model roster is stable this week (no additions; MAI-Code-1-Flash out), and the shipping emphasis has moved to enterprise control surfaces -- centrally managed agent permissions and sandboxes are what unblocks agent mode in regulated shops. **Still ahead on this page's calendar:** no earlier than 9/28 the unified Copilot experience defaults on; 10/01 upfront per-seat charging for existing card/PayPal customers; 10/02 the four-model deprecation slate.Source: GitHub Changelog (2026-09-10-mai-code-1-flash-deprecated; 2026-09-14-configure-cost-and-quality-in-copilot-auto-model-selection; 2026-09-09-enterprise-managed-permissions-for-github-copilot-agent-operations; 2026-09-08-enterprise-managed-sandbox-in-copilot-for-jetbrains; 2026-09-11-auto-resolution-and-analysis-updates-in-copilot-code-review; 2026-09-10-github-copilot-weekly-releases-september-7; 2026-09-11-add-vs-code-agents-to-copilot-usage-metrics; 2026-09-09-remediate-code-quality-findings-with-agentic-autofix) -- all fetched 2026-09-14 via curl · 2026-09-14
  • THREE FRONTIER MODELS LANDED IN COPILOT IN FOUR DAYS -- AND THE ENABLEMENT POSTURE SPLITS EXACTLY ALONG DATA-RETENTION LINES (2026-09-01 to 2026-09-04, vendor-primary): GitHub added **Claude Fable 5.1 (2026-09-01)**, **Gemini 3.8 Flash (2026-09-03)** and **GPT-6 Astra (2026-09-04)**, each within a day or two of its vendor launch. **THE IMPORTANT DIFFERENCE IS NOT THE MODELS, IT IS WHICH ONES TURN THEMSELVES ON.** For Astra and Gemini 3.8 Flash, GitHub's wording is identical and automatic: '**Under default model enablement, new models are enabled automatically unless an administrator has turned off the global default or explicitly disables this model.**' For Claude Fable 5.1, it is the opposite: '**Copilot Enterprise and Copilot Business plan administrators must enable the Claude Fable 5.1 policy in Copilot settings. The policy is off by default.**' **The stated reason is data retention, not capability:** '**Unlike other Claude models in GitHub Copilot, Claude Fable 5.1 requires data retention by default to operate Anthropic's safety classifiers.**' Zero data retention is available only 'for certain eligible enterprise customers'. **This confirms the dividing line this page identified on 2026-08-26: default-on enablement in enterprise Copilot is gated by licensing and data-retention posture, not by how good the model is.** Anthropic's newest and strongest model is the one that has to be switched on by hand, precisely because its safeguards require retaining data. **AVAILABILITY AND TIERS:** Astra and Fable 5.1 are **Pro+, Max, Business and Enterprise** (note: **not** Pro); Gemini 3.8 Flash reaches **Pro, Pro+, Max, Business and Enterprise**. All three appear across VS Code, Visual Studio, Copilot CLI, the coding/cloud agent, the Copilot app, JetBrains, Xcode and Eclipse; Astra and Fable 5.1 also on github.com and GitHub Mobile. Rollout is gradual in every case. **BILLING:** Astra and Fable 5.1 bill at **provider list pricing** under usage-based billing; Gemini 3.8 Flash bills at **introductory provider pricing through 2026-12-31** -- so the Copilot cost of that model doubles on 2027-01-01 along with Google's rate card. **GitHub's own testing note on Astra is about process rather than output: it 'plans and validates as it goes, batches diagnosis with verification, and independently confirms its results before declaring a task done.'**Source: GitHub Changelog (2026-09-04-gpt-6-astra-is-generally-available-in-github-copilot, 2026-09-03-gemini-3-8-flash-is-now-available-in-github-copilot, 2026-09-01-claude-fable-5-1-generally-available-in-github-copilot) -- all fetched 2026-09-05 · 2026-09-04
  • A FOUR-MODEL DEPRECATION SLATE DATED 2026-10-02 -- INCLUDING CLAUDE OPUS 4.7 AND TWO GEMINI FLASH VERSIONS (2026-09-03, vendor-primary): GitHub will remove the following '**across all GitHub Copilot experiences (including Copilot Chat, inline edits, ask and agent modes, and code completions) on October 2nd, 2026**': **Gemini 3.5 Flash** and **Gemini 3.6 Flash** (both -> Gemini 3.8 Flash), **Kimi K2.7 Code** (-> Kimi K3), and **Claude Opus 4.7** (-> Claude Opus 5). **READ THE GEMINI PAIR TOGETHER WITH THE 3.8 LAUNCH: Google shipped 3.8 Flash on 9/02 and GitHub scheduled the retirement of two older Flash versions the next day.** The Flash line is now turning over faster inside Copilot than most teams re-pin their model choices -- three Flash releases in six weeks upstream, two Flash retirements downstream. **THE ADMIN TRAP GITHUB FLAGS EXPLICITLY:** the suggested alternatives are not automatically usable. '**Copilot Enterprise and Copilot Business administrators may need to enable access to the alternative models through their model policies in Copilot settings.**' So an org that is not paying attention can lose Opus 4.7 on October 2 and find Opus 5 not enabled behind it. Verify the policy is on before the date, not after. **No action is required to remove the models after deprecation, and GitHub directs Enterprise customers with concerns to their account team.**Source: GitHub Changelog (github.blog/changelog/2026-09-03-upcoming-deprecation-of-selected-github-copilot-models/) -- fetched 2026-09-05 · 2026-09-03
  • COPILOT BUSINESS AND ENTERPRISE SIGNUPS REOPEN -- WITH UPFRONT PER-SEAT CHARGING THAT HITS EXISTING CUSTOMERS ON 2026-10-01 (2026-09-03, vendor-primary; the dated flip this page has tracked since 8/28, and it landed two days late): GitHub is '**gradually reopening sign-ups for Copilot Business and Copilot Enterprise customers paying by credit card or PayPal over the next couple of weeks**', alongside '**strengthening account vetting and updating billing experiences**' for those customers. **THE BILLING CHANGE IS THE PART THAT COSTS MONEY AND IT IS NOT LIMITED TO NEW CUSTOMERS.** Two changes apply to card/PayPal Business and Enterprise accounts: '**All new Copilot Business or Copilot Enterprise seat assignments will require payment for each seat before users gain Copilot access**', and '**At the start of the next billing cycle, all Copilot Business and Copilot Enterprise seats assigned will incur an upfront charge. This includes existing customers starting October 1, 2026.**' **So an established org that has been billed in arrears switches to paying for every assigned seat upfront on October 1** -- a working-capital change, not a price change. Included usage 'may be prorated across the month to align with the seat cost proration.' **WHAT IS EXPLICITLY NOT CHANGING:** plan pricing, seat prorations, and the ability to buy usage beyond included allowances. **One trap worth knowing: 'Fully canceling Copilot and returning later may trigger updated account vetting and billing behaviors'** -- so churning out and back in is not a way to keep the old billing treatment. Invoiced enterprise customers are outside the scope of this notice.Source: GitHub Changelog (github.blog/changelog/2026-09-03-reopening-copilot-business-and-enterprise-signups/) -- fetched 2026-09-05 · 2026-09-03
  • ENTERPRISE-MANAGED SETTINGS CAN NOW PIN ANY MODEL AS THE DEFAULT, INCLUDING PER-TEAM (2026-09-02, vendor-primary): GitHub made model defaults in enterprise-managed settings generally available -- '**You can now set your preferred GitHub Copilot model as the default for new conversations through enterprise-managed settings**', and '**customize the default by enterprise team, assigning a different default model based on team membership**' by marking the model key overridable and updating `team-mappings.json`. Users not covered by a team mapping inherit the enterprise default. **WHY THIS MATTERS THIS WEEK SPECIFICALLY:** it lands in the same window as three model additions and a four-model deprecation slate, and it is the control that makes those manageable -- an org can pin Gemini 3.8 Flash for a cost-sensitive team and Fable 5.1 or Astra for a team doing long-horizon work, rather than accepting one global default. **It also partially offsets the default-enablement asymmetry above:** Fable 5.1 will not enable itself, but once an admin turns the policy on, it can be made a team's default rather than something each developer selects. GA on **Copilot Business and Copilot Enterprise**, in the GitHub Copilot app, Copilot CLI and Visual Studio Code.Source: GitHub Changelog (github.blog/changelog/2026-09-02-enterprise-managed-settings-support-any-default-model/) -- fetched 2026-09-05 · 2026-09-02
  • COPILOT CODE REVIEW CAN NOW APPROVE PULL REQUESTS, AND CONTENT EXCLUSIONS REACHED GA IN THE APP AND CLI (2026-09-01 / 2026-09-02, vendor-primary): two smaller changes that matter for governance. **Copilot code review can now approve pull requests (2026-09-01)** -- following the 2026-08-27 removal of the 300-file / 20,000-LOC cap, the reviewer has gone from commenting to being able to register an approval. **Treat this as a branch-protection question, not a feature:** if your repository counts Copilot's approval toward a required-reviewers rule, an automated reviewer can now satisfy a human gate. Check the setting before assuming it cannot. **Content exclusions are generally available in the Copilot app and Copilot CLI (2026-09-02)**, closing a real gap -- exclusions that already applied in the IDE now also apply on those two surfaces, so files you had excluded from Copilot in VS Code are no longer reachable through the CLI. **The standing double-billing caveat on code review still applies.**Source: GitHub Changelog (2026-09-01-copilot-code-review-can-now-approve-pull-requests, 2026-09-02-content-exclusions-generally-available-in-copilot-app-and-cli) -- fetched 2026-09-05 · 2026-09-02
  • THE COPILOT CODE REVIEW SIZE CAP IS GONE, AND IT NOW REVIEWS BOT-AUTHORED PRs (2026-08-27, vendor changelog): three changes shipped together, and the first is the one that actually changes what the product can do. **(1) THE 300-FILE / 20,000-LINE LIMIT HAS BEEN REMOVED.** Copilot code review previously refused pull requests above **300 files or 20,000 lines of code**; that ceiling is lifted, so large refactors, dependency sweeps and generated-code PRs -- exactly the reviews a human is least willing to do carefully -- are now in scope for the first time. **(2) BOT-AUTHORED PULL REQUESTS ARE NOW REVIEWED**, including PRs opened by **Copilot cloud agent**, when reviews are requested automatically. GitHub is explicit that the cloud agent's PRs now get '**a full agentic review**' rather than the previously limited experience. **This closes a real loop: Copilot's own agent writes the PR, and Copilot reviews it.** Whether you regard that as useful automation or as marking your own homework is a judgement call the changelog does not make for you, and there is no second-model review option. Billing detail worth catching: organizations with the '**Allow members without a Copilot license to use Copilot code review**' policy enabled can **bill that usage directly to the organization**. **(3) RESOLUTION REASONS:** when you resolve a Copilot comment you can now pick **Addressed**, **Won't fix**, or **Incorrect** from a dropdown beside the resolve button -- the first structured signal GitHub collects on whether its review comments were actually right, and the 'Incorrect' option is the one to use if you want that feedback to mean anything. **THE COST WARNING ON THIS PAGE APPLIES DOUBLE NOW:** since 2026-06-01 Copilot code review bills **twice** -- the premium request plus GitHub Actions runtime -- so removing the size cap removes the thing that was implicitly capping your spend on runaway PRs. **A 5,000-file migration PR is now reviewable, and now billable.** The changelog does not state which plans or tiers these three changes apply to.Source: GitHub Changelog (github.blog/changelog/2026-08-27-copilot-code-review-resolution-reasons-and-expanded-capabilities/, dated 2026-08-27) -- fetched 2026-08-31 · 2026-08-27
  • COPILOT MOVES INTO SLACK AND MICROSOFT TEAMS -- AGENT SESSIONS BECOME MULTIPLAYER (2026-08-21, vendor-primary): GitHub shipped two chat-surface integrations in **public preview** on the same day. **Slack:** the GitHub integration now brings '**the agentic capabilities of GitHub Copilot CLI and the GitHub Copilot app into Slack**'. Mention **@GitHub** in a DM, channel or thread to start an agent session that can answer questions about code and GitHub activity, triage bug reports, create/update/label issues, and investigate failures then implement and validate changes. **GitHub is also a launch partner for Slack Code, a new Slack channel type designed for agents** -- worth watching as a category, not just a Copilot feature. **Microsoft Teams:** mention **@GitHub** to turn a discussion into a **shared Copilot cloud agent session that everyone in the conversation can see and steer**; anyone can ask questions and add context, but **only participants with write access to the repo can trigger Copilot to make changes** -- the governance boundary that makes this safe to enable. GitHub's framing is meeting-to-work: hand an action item to Copilot during standup and have it investigating before the meeting ends. **THE STRATEGIC POINT FOR RANKINGS: Copilot's agent is no longer editor-bound, and 'shared, observable agent session' is a collaboration primitive no editor-only assistant currently matches.** Both are previews, so treat durability as unproven.Source: GitHub Changelog (github.blog/changelog/2026-08-21-the-new-github-copilot-experience-in-slack/ and github.blog/changelog/2026-08-21-shared-agentic-work-with-github-copilot-in-microsoft-teams/) -- both fetched 2026-08-28 via curl · 2026-08-21
  • COPILOT HARNESS GA IN JETBRAINS, PLUS A BUILT-IN JETBRAINS MCP SERVER (2026-08-24, vendor-primary): **Copilot harness reached general availability in Copilot for JetBrains**, which GitHub credits with faster feature delivery and better code quality. Shipping alongside: **built-in JetBrains MCP server support in public preview**, exposing IDE capabilities to MCP-driven agent workflows **without requiring an external server setup** -- the friction that has kept most IDE MCP usage to early adopters. Also **/review harness integrations directly inside JetBrains**, and a What's New tab after plugin updates. **Read alongside the 2026-08-18 enterprise-managed-settings release for JetBrains: GitHub is closing the gap between its VS Code flagship and the JetBrains plugin on both agent capability and enterprise governance in the same fortnight.**Source: GitHub Changelog (github.blog/changelog/2026-08-24-copilot-harness-generally-available-in-copilot-for-jetbrains/) -- fetched 2026-08-28 via curl · 2026-08-24
  • GLOBAL MODEL POLICY GOES GA -- AND THE 8/17 WARNING NOT TO ASSUME IT AUTO-ENABLES EVERYTHING WAS CORRECT (2026-08-26, vendor-primary): The default model policy announced in July for Copilot Business/Enterprise began **gradual enforcement on 2026-08-26, rolling out through 2026-09-01**, so it takes effect at different times per enterprise. **We flagged on 2026-08-17 that this date should not be assumed to auto-enable Gemini 3.7 Flash and Grok 4.6. That caution is now confirmed, and the exclusion rule is broader and more interesting than model-by-model defaults:** GitHub excludes from default enablement (a) **all open-weight models -- it names DeepSeek and Kimi K2** and (b) **any model not covered by GitHub's data retention agreement -- it names Fable 5**, regardless of what an admin's policy is set to. **So the dividing line is licensing and data-retention posture, not capability or vendor prestige** -- an Anthropic frontier model and an open-weight Chinese model are excluded for different reasons but land in the same bucket. Unconfigured and newly-GA models inherit a **'Delegate to default policy'** state, a live setting that tracks the policy rather than freezing a value. **Explicit admin choices are preserved** -- deliberately enabled or disabled models are untouched. Four resulting states: Enabled, Disabled, Delegate to enterprise teams/apps or organizations, Delegate to default policy. **GitHub is consulting on removing 'Delegate to default policy' later** so every policy is an explicit decision.Source: GitHub Changelog (github.blog/changelog/2026-08-26-global-model-policy-generally-available/, on-page 'August 26, 2026') -- fetched 2026-08-28 via curl · 2026-08-26
  • COPILOT BUSINESS/ENTERPRISE SIGNUPS REOPEN 9/1 -- BUT SEATS MUST NOW BE PAID FOR BEFORE THEY WORK (2026-08-28, vendor-primary): GitHub published a slate of upcoming policy and billing changes. **The headline is that self-serve signups had been closed and are coming back: 'Starting September 1, 2026, GitHub will start reenabling sign-ups for new Copilot Business and Copilot Enterprise customers paying by credit card or PayPal,' with 'strengthened account vetting'** -- an anti-abuse measure, and an implicit admission the previous flow was being abused. **THE CHANGE THAT COSTS EXISTING CUSTOMERS MONEY IS THE SECOND ONE, AND IT IS NOT A PRICE RISE:** all new Business/Enterprise seat assignments will **require payment for each seat before users gain Copilot access**, and at the start of the next billing cycle **all assigned seats incur an upfront charge**. For existing credit-card/PayPal customers this begins **2026-10-01**. GitHub states plainly '**Copilot Business and Copilot Enterprise prices are not changing**' -- so this is a **cash-flow and provisioning change, not a rate change**, and anyone reading only the headline will get that backwards. **Retained terms:** mid-cycle seats still prorate; included usage may be prorated to match seat proration; overage still purchasable; spend controls, usage tracking and AI credits unchanged. **Retained sting: revoking a seat still yields no prorated refund**, which matters more now that seats are charged upfront. **SEPARATE DATED ITEM -- 'no earlier than September 28, 2026': Copilot cloud agent, Copilot Chat on github.com and Copilot Chat in GitHub Mobile converge into a single unified Copilot experience under one policy, enabled by default after launch**, with cloud agent moving onto Sandbox. Three separate admin policies collapse into one, and it defaults on -- worth flagging to anyone who had deliberately disabled just one of the three.Source: GitHub Changelog (github.blog/changelog/2026-08-28-upcoming-changes-to-github-copilot-policies-and-billing/) -- fetched 2026-08-28 via curl · 2026-08-28
  • ENTERPRISE MANAGED SETTINGS LAND IN COPILOT FOR JETBRAINS -- INCLUDING AN MCP ALLOWLIST AND A KILL SWITCH FOR AUTOPILOT (2026-08-18, vendor changelog): admin governance that had been VS Code-shaped now reaches JetBrains IDEs, and one of the four controls is more consequential than the changelog's flat tone suggests. **THE FOUR CONTROLS:** (1) **Plugin governance** -- `enabledPlugins` to force a plugin on or off, `extraKnownMarketplaces` to approve additional sources, `strictKnownMarketplaces` to limit installation to approved sources only. (2) **MCP server allowlist** -- `allowedMcpServers` and `deniedMcpServers` centrally control which MCP servers developers may connect to, which GitHub says '**prevents connections to servers outside the enterprise allowlist**'. (3) **Managed OpenTelemetry** -- admins set the collector endpoint, protocol, service name, resource attributes and content-capture policy, and '**managed values take precedence over developer settings**'. (4) **Organization-controlled permission modes** -- `permissions.disableBypassPermissionsMode` **prevents the Copilot agent in JetBrains from using Bypass Approvals or Autopilot**. **WHY (2) AND (4) ARE THE ONES TO CARE ABOUT:** an MCP server is arbitrary third-party code with tool access inside a developer's editor, and until now a JetBrains shop had no central way to bound which ones were reachable -- this closes a real supply-chain gap rather than adding a convenience toggle. And `disableBypassPermissionsMode` is the first documented way for an org to stop Copilot's agent from running unattended in JetBrains, which is the control a security review actually asks for. **SCOPE:** applies to enterprises on a Copilot plan, via the enterprise managed settings reference; requires the latest JetBrains plugin. No pricing changeSource: GitHub Changelog (github.blog/changelog/2026-08-18-enterprise-managed-settings-in-github-copilot-for-jetbrains/, date-prefixed slug, fetched 2026-08-20) · 2026-08-18
  • MAI-CODE-1.1-FLASH ARRIVES AND MAI-CODE-1-FLASH IS ALREADY SCHEDULED TO DIE (both 2026-08-11, GitHub changelog) -- A ONE-MONTH FORCED MIGRATION. **(1) The new model.** MAI-Code-1.1-Flash, 'Microsoft's latest small-tier coding model', is rolling out in Copilot with **native vision support for image understanding**, improved coding quality and tool use. **Entitlement differs by plan in a way that matters: Free and Student users get it by automatic selection, while Pro, Pro+, Max, Business and Enterprise get it as a manual selection.** It carries a **0.25x premium request multiplier** for annual subscribers -- the cheapest multiplier in the picker -- and GitHub states a **73% lower list price than MAI-Code-1-Flash**, the model it replaces. **For Business and Enterprise the policy is OFF BY DEFAULT**, so an admin has to enable it. **(2) The retirement.** The same day, GitHub set **MAI-Code-1-Flash to retire on 2026-09-10**, across 'all GitHub Copilot experiences', individual and enterprise, with MAI-Code-1.1-Flash named as the suggested alternative. **The practical trap:** the replacement is off by default for orgs while the outgoing model has a hard date, so an Enterprise admin who does nothing leaves users with the old model until 9/10 and then with neither. Enable the 1.1 policy before the deadline. GitHub says 'no action is required to remove the older model once it has been deprecated' but does not spell out what happens to in-flight requests after the cutoff. **Note also that 9/10 sits after the 2026-08-26 auto-enablement change** -- MAI-Code-1.1-Flash is a first-party GA model, so it is the kind of model that policy is designed to auto-enable, but GitHub's own changelog still describes it as off by default today; do not assume the 8/26 flip retroactively covers itSource: GitHub changelog (github.blog/changelog/2026-08-11-mai-code-1-1-flash-available-in-github-copilot/ and github.blog/changelog/2026-08-11-upcoming-deprecation-of-mai-code-1-flash/, both fetched 2026-08-13) · 2026-08-11
  • COPILOT MEMORY + OLLAMA BYOK REACH JETBRAINS, AND PER-MODEL TOKEN BREAKDOWN LANDS IN BILLING (2026-08-11, GitHub changelog): **(1) Copilot memory in JetBrains** -- 'Copilot memory can now retain and recall useful information across agent chat sessions', so project details and preferences persist between conversations rather than being re-fed each time. **(2) Ollama as a BYOK provider in JetBrains** -- 'You can now use Ollama as a BYOK provider in GitHub Copilot for JetBrains', which means **local models running on your own machine inside Copilot's JetBrains client**. That is the meaningful one for anyone with data-residency constraints or a preference for open weights: it is a route to using Copilot's interface without sending code to a hosted frontier model. GitHub's changelog does **not** state which plans or SKUs these require -- recorded as unspecified rather than assumed to be all-plans. **(3) Per-model token breakdown in the usage report** -- 'You can now see a per-model breakdown of the tokens behind your AI credits in the usage report', showing **input, output, cache read and cache write tokens per model** alongside credits consumed, on **Business, Enterprise and Individuals**. This continues the metering-transparency thread that has been the most useful Copilot change since the 6/01 usage-billing cutover: cache read/write visibility in particular is what lets you tell whether a high bill is real work or cache missesSource: GitHub changelog (github.blog/changelog/2026-08-11-copilot-memory-and-ollama-in-github-copilot-for-jetbrains/ and github.blog/changelog/2026-08-11-per-model-token-breakdown-in-the-usage-report/, both fetched 2026-08-13) · 2026-08-11
  • AGENT PLUGINS 1.0 -- A VENDOR-NEUTRAL PACKAGING STANDARD FOR SKILLS + MCP SERVERS (announced 2026-08-12, spec published 2026-08-06, GitHub changelog): GitHub describes it as **'an open standard that packages agent skills and MCP servers into one installable plugin that is governed independently of any single vendor.'** The problem it solves is duplication: plugins previously needed **separate manifests and directory layouts for each client**, so one plugin now works across multiple agent clients instead of shipping vendor-specific versions. **Available in VS Code, Copilot CLI, the GitHub Copilot SDK and the GitHub Copilot app**, and **generally available on all Copilot plans**. **Not a breaking change** -- 'Existing GitHub Copilot plugins that don't target Agent Plugins 1.0 remain supported, with no migration required.' **Why it is worth watching beyond Copilot:** the 'governed independently of any single vendor' framing puts this in the same category as MCP itself, which GitHub also did not own but adopted -- if other agent clients implement it, plugin portability stops being a Copilot feature and becomes an ecosystem property. Whether that happens is not yet evidenced; today it is GitHub-published and GitHub-implementedSource: GitHub changelog (github.blog/changelog/2026-08-12-agent-plugins-1-0-in-vs-code-copilot-cli-and-the-copilot-app/, fetched 2026-08-13) · 2026-08-12
  • TWO FRONTIER MODELS LANDED IN THE PICKER IN 48 HOURS -- GEMINI 3.7 FLASH (8/13) AND GROK 4.6 (8/14), BOTH ON EVERY PAID SKU (vendor changelog, both fetched 2026-08-17): **(1) Gemini 3.7 Flash, 2026-08-13** -- GitHub added Google's newest Flash model **the same day Google launched it**, citing improvements in web/app development and agentic coding workflows. Availability: '**Copilot Pro, Pro+, Max, Business, and Enterprise users**' -- the broad list, including base Pro. **(2) Grok 4.6, 2026-08-14** -- xAI's latest coding model, which GitHub says '**performed especially well on longer-horizon tasks requiring sustained reasoning and tool use, making it a strong fit for complex coding workflows**.' Same SKU list: '**Copilot Pro, Pro+, Max, Business, and Enterprise SKUs**.' xAI published its own post the same day confirming it is 'now live in GitHub Copilot,' and noting the model is also available direct from the SpaceXAI console at **$2 per million input / $6 per million output** -- unchanged from Grok 4.5, so this is a capability bump at a flat price. **Surfaces for both:** VS Code, Visual Studio, Copilot CLI, the Copilot cloud agent, the GitHub Copilot app, JetBrains, Xcode and Eclipse. **ORG ADMINS -- BOTH ARE STILL OFF BY DEFAULT.** For Gemini 3.7 Flash, '**Copilot Enterprise and Copilot Business plan administrators must enable the Gemini 3.7 Flash Preview policy in Copilot settings before anyone in their organization can select it**' (note GitHub labels the policy **Preview**). For Grok 4.6, '**the policy is off by default**' for Business and Enterprise admins. xAI words it more loosely -- 'some businesses and enterprises will need to enable the model from Copilot settings' -- but GitHub's changelog is the authoritative one for org posture. **HOW THIS INTERACTS WITH THE 2026-08-26 DEFAULT-ENABLEMENT FLIP, because it is easy to get wrong:** the 8/26 change auto-enables newly released GA models covered by GitHub's data-retention agreement. Neither of these is automatically covered by that -- **Gemini 3.7 Flash is carrying a Preview policy label**, and both shipped **before** 8/26 with an explicit off-by-default posture, so **do not assume 8/26 retroactively switches them on**. If you want either model available to your org, enable it manually. **NO MULTIPLIER PUBLISHED FOR EITHER** -- GitHub's changelog states no premium-request multiplier and no context-window figure for Gemini 3.7 Flash or Grok 4.6, so any multiplier you see quoted for these two in an aggregator is not from the changelog. Both bill usage-based at provider list pricing, consistent with the Opus 5, Grok 4.5 and Kimi K3 adds. **Net:** the picker now spans five frontier vendors, and the two newest additions both arrived on day one or day two of their upstream launch -- Copilot's model-freshness lag is effectively goneSource: GitHub changelog (github.blog/changelog/2026-08-13-gemini-3-7-flash-is-now-available-in-github-copilot and github.blog/changelog/2026-08-14-grok-4-6-is-now-available-in-github-copilot) plus SpaceXAI (x.ai/news/grok-4-6-github-copilot, on-page date Aug 14, 2026) -- all fetched 2026-08-17 · 2026-08-14
  • SPEND VISIBILITY FINALLY LANDS IN THE CHAT ITSELF (2026-08-10, vendor changelog) -- the single most useful change for anyone on usage-based billing: Copilot on the web (github.com/copilot) added **token spend indicators in chat**, giving you '**per-session and per-message quota**' and, in GitHub's words, 'greater visibility into your token spend and budget.' **Why this is more than a UI tweak on this page specifically:** since the 2026-06-01 billing cutover, every Copilot page here has carried the same structural complaint -- premium requests and credits burn in ways that are hard to see until the bill arrives, and the 8/03 reasoning-level control made that worse by adding a knob that 'consumes more tokens, and therefore more credits.' Per-message spend in the chat is the first in-context feedback loop on that. Availability: **all Copilot plans**. Shipped alongside the ability to **minimize a conversation and resume it** -- browse GitHub while a response is generating, then pick recent conversations back up. **A related transparency change shipped 2026-08-07 in the Copilot app:** '**Auto now shows which model handled each completed request, plus AI credit and cache details when they're available**' -- so Auto-mode routing, which previously gave you no way to know what you were billed for, is now itemised. Read together, these two are GitHub closing the metering-opacity gap rather than adding capabilitySource: GitHub changelog (github.blog/changelog/2026-08-10-copilot-on-web-expands-conversation-controls/ and github.blog/changelog/2026-08-07-github-copilot-weekly-releases-august-3/, both fetched 2026-08-10) · 2026-08-10
  • COPILOT CODE REVIEW EFFORT LEVELS GO GA, AND THE NAMES CHANGED (2026-08-07, vendor changelog): the **Lite** and **Balanced** effort levels for Copilot code review moved from public preview to **general availability** -- and note the rename, because it will break your muscle memory and any internal docs: **these were previously called Low and Medium**. You pick a level **per review** (Lite for routine changes, Balanced for complex or sensitive code), and **organization admins can set a default effort level across all repositories**. Timeline events and pull-request comments now display which effort level was applied, so reviewers can see whether a given review was a light pass or a deep one. Availability: **Copilot Pro, Pro+, Max, Business, and Enterprise** -- i.e. every paid tier, not the free tier. No additional cost was announced beyond existing subscriptions, but remember the standing gotcha on this page: since 2026-06-01 Copilot code review bills **twice** -- the premium request plus GitHub Actions runtime -- so a default of Balanced across every repo is a spend decision, not just a quality oneSource: GitHub changelog (github.blog/changelog/2026-08-07-copilot-code-review-effort-levels-are-generally-available/, fetched 2026-08-10) · 2026-08-07
  • KIMI K3 IS GA IN COPILOT -- AND IT LANDS ON EVERY PAID SKU INCLUDING BASE PRO (2026-08-06, vendor changelog): GitHub made **Kimi K3**, Moonshot's 2.8T-parameter open-weight model, generally available in the Copilot picker. GitHub's framing, verbatim: the model '**shows frontier-level abilities on agentic coding with highly cost-effective pricing**'. **HOSTING MATTERS HERE: Kimi K3 is hosted by GitHub on Fireworks AI**, not called out to Moonshot's own API -- so your prompts go to a US inference provider under GitHub's arrangement rather than to Moonshot. Billing is **provider list pricing under usage-based billing**, the same posture as the Opus 5 and Grok 4.5 adds. **AVAILABILITY IS BROADER THAN THE RECENT FRONTIER ADDS:** K3 is rolling out to **Pro, Pro+, Max, Business and Enterprise** -- note that includes **base Pro**, where Claude Opus 5 (added 7/24) was Pro+/Max/Business/Enterprise only. Surfaces: VS Code, Visual Studio, Copilot CLI, Copilot cloud agent, the GitHub Copilot app, github.com, GitHub Mobile on iOS and Android, JetBrains, Xcode and Eclipse. Rollout is gradual, so absence from your picker is not a fault. **ORG ADMINS -- THIS ONE DOES NOT AUTO-ENABLE: K3 is off by default for Copilot Business and Copilot Enterprise**, and 'plan administrators must enable the Kimi K3 policy in Copilot settings before anyone in their organization can select it.' That is consistent with, not an exception to, the 2026-08-26 default-model-enablement change described below -- **open-weight models are explicitly excluded from auto-enablement**, so K3 will still need a manual switch after 8/26. GitHub adds an unusually direct caution: 'we recommend administrators review open-weight models against their own security, compliance, and data-governance requirements before enabling them.' Read: the picker now carries two Moonshot generations (K2.7-Code and K3), and the cheapest credible frontier-class coding option in Copilot is an open-weight Chinese model -- which is exactly the decision GitHub is pushing onto compliance teamsSource: GitHub changelog (github.blog/changelog/2026-08-06-kimi-k3-is-now-available-in-github-copilot/, fetched 2026-08-06) · 2026-08-06
  • GITHUB SPARK IS BEING RETIRED -- CLOSED TO NEW USERS 2026-08-04, EXPORT BY 2026-08-31 (vendor changelog): GitHub is winding down **Spark**, its natural-language app-builder on github.com. Verbatim: 'Beginning August 4, 2026, GitHub Spark no longer accepts new users or allows the creation of new apps.' **Existing apps you already deployed keep working after shutdown** -- but if you want to keep *editing* an app you must **export the code before 2026-08-31** (open the Spark workbench, select '...', then 'Create repository'). The sting for anyone who built on it: Spark's AI features ran on **GitHub Models, which was retired 2026-07-30**, so any app calling the `llm()` function **already stopped working on 7/30** -- to restore AI behaviour you must swap in your own inference provider, supply your own API key, and take on the billing yourself. Read this as the second half of one decision, not two: GitHub retired the free inference service and then closed the product that depended on it, both inside five days. If you are evaluating Copilot, note that Spark was a separate product -- **this does not affect Copilot subscriptions or the model picker**Source: GitHub changelog (github.blog/changelog/2026-08-04-upcoming-deprecation-of-github-spark-on-github-com/) · 2026-08-04
  • COPILOT CLOUD AGENT GETS A REASONING-LEVEL CONTROL (2026-08-03, vendor changelog): when you delegate a task to Copilot cloud agent you can now **pick a reasoning level alongside the model**, controlling 'how much the model reasons before it responds,' and the agent uses that level for the whole run. It applies only to 'models that support it' -- GitHub did not publish which models, which named levels, or the default. **Billing matters here:** GitHub's own wording is that a higher level 'can improve answers to complex problems, but it consumes more tokens, and therefore more credits' -- so this is a knob that can quietly raise your premium-request burn on a metered plan. Availability: **all paid plans that include Copilot cloud agent -- Pro, Pro+, Business, Enterprise, and Max** (not the free tier). Same-week companion change: the **Copilot billing preview app was retired 2026-08-04**Source: GitHub changelog (github.blog/changelog/2026-08-03-customize-the-reasoning-level-for-copilot-cloud-agent/, github.blog/changelog/2026-08-04-retiring-the-copilot-billing-preview-app/) · 2026-08-03
  • GITHUB MODELS IS GONE -- RETIREMENT EXECUTED 2026-07-30 (vendor-confirmed, runtime-confirmed): the free model playground and catalog are dead, with **no grandfathering and no read-only window**. GitHub's changelog, verbatim: 'As of July 30, 2026, GitHub Models is now retired. The playground, model catalog, inference API, and bring your own key (BYOK) are no longer available to any customer, including existing customers with active usage.' The `models.github.ai` catalog endpoint now returns **HTTP 410 Gone**. IMPORTANT: **the free playground inference was not migrated anywhere -- it simply ended.** GitHub points users to Microsoft/Azure AI Foundry (the changelog says 'Microsoft Foundry', the docs say 'Azure AI Foundry' -- same product, inconsistent naming) or to Copilot itself; **both are paid**. Also note GitHub's own docs stress that 'GitHub Models was a separate service from GitHub Copilot and is unrelated to GitHub Copilot services' -- so this does not affect your Copilot subscription or model pickerSource: GitHub changelog (github.blog/changelog/2026-07-30-github-models-is-now-retired/), GitHub docs (docs.github.com/en/github-models), live probe of models.github.ai returning 410 · 2026-07-30
  • MODEL DEPRECATIONS RESOLVED + A NEW SEPT 1 SLATE (2026-07-31, vendor changelog -- this supersedes the earlier 'removal date not yet specified' note): (1) **Gemini 2.5 Pro and Gemini 3 Flash were actually deprecated on 2026-07-31**, verbatim 'As of today, July 31, 2026, we have deprecated the following models across all GitHub Copilot experiences.' Replacements: **Gemini 2.5 Pro -> Gemini 3.1 Pro (Preview)** and **Gemini 3 Flash -> Gemini 3.6 Flash**. (2) Announced the same day, a further slate deprecates on **2026-09-01**: **Gemini 3.1 Pro -> Gemini 3.6 Flash** (note the awkward consequence -- 3.1 Pro is the *replacement* for 2.5 Pro on 7/31 and is itself deprecated a month later, so migrating 2.5 Pro users to 3.1 Pro buys you about four weeks); **Claude Opus 4.5 and Opus 4.6 -> 'Claude Opus 4.7, Claude Opus 4.8, or Claude Opus 5'**; **Claude Sonnet 4.5 -> Sonnet 5**; **Claude Sonnet 4.6 -> Sonnet 5**, with the carve-out that 'Claude Sonnet 4.6 will remain available to individual GitHub Copilot subscribers on annual plans so they retain a Sonnet offering'; and **Raptor Mini -> MAI-Code-1-Flash**. GitHub's instruction: 'update your workflows and integrations to use the supported model before these dates'Source: GitHub changelog (github.blog/changelog/2026-07-31-gemini-2-5-pro-and-gemini-3-flash-deprecated/ and /2026-07-31-upcoming-august-2026-model-deprecations-in-github-copilot/, both fetched 2026-08-03) · 2026-07-31
  • ORG MODEL DEFAULTS FLIP ON 2026-08-26 -- ACT BEFORE THEN IF YOU WANT THE OLD BEHAVIOUR (announced 2026-07-29, vendor changelog): GitHub is reversing the admin-opt-in posture for new models. Under the new **default model enablement** policy, newly released Copilot models that are generally available **and covered by GitHub's data retention agreement become available to Business and Enterprise users automatically**, instead of requiring an admin to switch each one on. **Excluded from auto-enablement: open-weight models (GitHub names DeepSeek and Kimi K2.7) and models not covered by GitHub's data retention agreement (GitHub names Fable 5)** -- those still require explicit enablement. Timing matters: **the policy is configurable now but has no effect until 2026-08-26**, and organisations that want to keep manual control must **disable the policy before that date**. This directly changes the 'models are off by default for orgs' guidance that applied to the Opus 5 and Grok 4.5 adds belowSource: GitHub changelog (github.blog/changelog/2026-07-29-default-model-enablement-for-copilot-business-and-enterprise/, fetched 2026-08-03) · 2026-07-29
  • COPILOT BILLING PREVIEW APP RETIRED TODAY, 2026-08-03 (vendor-confirmed): the standalone app 'built to help you understand your bill as Copilot moved to usage-based billing' is gone. Its replacement is built into GitHub: the **AI usage page in billing settings**, which GitHub says now covers things the app never could, including 'user-level budgets, cost centers, and how usage pools are allocated.' If you had the app installed for spend tracking, that workflow moves to billing settingsSource: GitHub changelog (github.blog/changelog/2026-07-07-copilot-billing-preview-app-will-be-retired-on-august-3/, fetched 2026-08-03) · 2026-08-03
  • MODEL ADD -- GROK 4.5 (2026-07-28): GitHub added xAI's Grok 4.5 to Copilot for **'Copilot Pro, Pro+, Max, Business, and Enterprise SKUs'** -- a wider plan list than Opus 5 got four days earlier, since it includes base Pro. Specs from the changelog: **up to 500,000 tokens of context**, text and image input, and selectable **low / medium / high reasoning effort**, positioned for fast agentic coding and complex multi-step workflows. Billing is usage-based at provider list pricing. **Business and Enterprise administrators must enable the Grok 4.5 policy in Copilot settings -- it is disabled by default**, so it will not simply appear for org usersSource: GitHub changelog (github.blog/changelog/2026-07-28-grok-4-5-is-now-available-in-github-copilot/, fetched 2026-07-29) · 2026-07-28
  • MODEL ADD -- CLAUDE OPUS 5 (2026-07-24): Copilot picked up Anthropic's new flagship **the same day Anthropic launched it**. Availability is **'Copilot Pro+, Max, Business, and Enterprise users'** -- note this **excludes base Copilot Pro**, unlike the Grok 4.5 add four days later. Billed usage-based at provider API list rates (GitHub published no multiplier and no context figure in the changelog), and **Enterprise/Business admins must enable the Claude Opus 5 policy** before their developers see it. Rollout was gradual. Practical read: the strongest Anthropic model is now a Copilot option, which narrows the model-access gap that pushed heavy users toward Cursor and Claude CodeSource: GitHub changelog (github.blog/changelog/2026-07-24-claude-opus-5-is-now-available-in-github-copilot/, fetched 2026-07-29) · 2026-07-24
  • JULY WEEK-2 CLUSTER (2026-07-01 → 2026-07-09, all vendor changelog): (1) **GPT-5.6 Sol, Terra, and Luna landed in Copilot 7/9** -- day-one availability with OpenAI's GA; the picker now spans GPT-5.6 (three tiers), Sonnet 5, Fable 5, and Kimi K2.7. (2) **Deprecation notice 7/2: Gemini 2.5 Pro and Gemini 3 Flash are being removed from Copilot** -- RESOLVED, the removal landed 2026-07-31; see the deprecation entry above for the replacements and the follow-on Sept 1 slate. (3) 7/7: **Kimi K2.7 extended to Business/Enterprise**, **Codex now works as an agent provider inside JetBrains IDEs**, the **GitHub Copilot app is available to all**, and the Copilot Billing Preview app retires 8/3 (DONE -- see the 2026-08-03 entry above). (4) 7/1: **Copilot vision GA** (images/PDFs in chat) + CLI auto model selection. Net: Copilot's picker is now the fastest-moving multi-vendor surface in the market -- three frontier vendors shipped into it inside 10 daysSource: GitHub Changelog (github.blog/changelog/label/copilot -- 2026-07-01 through 2026-07-09 entries) · 2026-07-09
  • LATE-JUNE/JULY CLUSTER (2026-06-30 → 2026-07-02, vendor changelog): (1) **Claude Sonnet 5 is GA in GitHub Copilot** (6/30) -- day-one availability alongside Anthropic's launch; the new 'most agentic Sonnet' model joins the picker. (2) **Browser tools for Copilot agents are GA in VS Code** (7/1) -- agents can drive a browser to test/verify their own web changes. (3) **GitHub Models is being fully retired on 2026-07-30** (7/1 announcement) -- the model-catalog/playground product winds down; migrate off before the date. (4) **Kimi K2.7 Code is GA in Copilot** (7/1) -- the first open-weight Chinese model in the picker; a lower-cost option that meters against the 6/1 credit system. Sonnet 5 note: available to Pro/Pro+/Max/Business/Enterprise, billed at provider list pricing under usage-based billing (no special multiplier), Zero Data Retention, rolling out across VS Code / Visual Studio / CLI / JetBrains / Xcode / Eclipse / mobile.Source: GitHub Changelog (github.blog/changelog/2026-06-30-claude-sonnet-5-is-generally-available-for-github-copilot/, .../2026-07-01-browser-tools-for-github-copilot-in-vs-code-are-generally-available/, .../2026-07-01-github-models-is-being-fully-retired-on-july-30-2026/, Kimi K2.7 Code GA entry 2026-07-01) · 2026-07-01
  • MODEL ADD (2026-06-09, vendor-primary github.blog/changelog): **Claude Fable 5 is generally available in GitHub Copilot** on launch day -- Anthropic's first public Mythos-class model lands in the Copilot model picker the same day it ships, ahead of Cursor (which still lacks it as of 6/10). For Copilot users this is currently the strongest available model on agentic coding benchmarks (vendor: SWE-Bench Pro 80.3%). Usage meters against the 6/1 credit system like other premium modelsSource: GitHub blog (github.blog/changelog/2026-06-09-claude-fable-5-is-generally-available-for-github-copilot/) · 2026-06-09
  • BILLING CUTOVER WENT LIVE (2026-06-01, vendor-primary github.blog/changelog): usage-based billing is now live for all GitHub Copilot users. Each plan ships with monthly included usage; beyond it, requests meter against GitHub AI Credits / token rates. Code completions and Next Edit suggestions stay free. **New Copilot Max tier introduced the same day** -- an upgrade for existing Student/Pro/Pro+ subscribers with higher included usage + higher spending limits. **New signups remain PAUSED for Student, Pro, Pro+, AND the new Max plan** ('we'll reopen sign-ups in the coming weeks'). Copilot code review now also consumes GitHub Actions minutes as of 6/1. Net: the 4/27-announced transition is no longer 'upcoming' -- it is the current state.Source: GitHub Changelog (github.blog/changelog/2026-06-01-updates-to-github-copilot-billing-and-plans/), GitHub Blog: moving to usage-based billing · 2026-06-01
  • MICROSOFT BUILD (2026-06-02): Microsoft's own **MAI-Code-1-Flash** -- a lightweight in-house coding model -- is rolling out to GitHub Copilot individual users in VS Code, in the model picker and under the default auto-picker. Vendor benchmarks vs Claude Haiku 4.5: SWE-Bench Pro 51.2% vs 35.2% (a 16-point lead), solves SWE-Bench Verified problems with up to 60% fewer tokens, and +28.9 on IF Bench (instruction-following). Microsoft frames it as 'better price-to-performance across coding benchmarks.' Significant because it is the first time a first-party Microsoft model is the cost-efficient default inside Copilot rather than an OpenAI or Anthropic model -- part of the broader 'seven new MAI models' Build wave that also includes MAI-Thinking-1 (reasoning), MAI-Image-2.5, MAI-Transcribe-1.5, and MAI-Voice-2.Source: Microsoft AI (microsoft.ai/news/introducingmai-code-1-flash/), Microsoft Build 2026 newsroom (news.microsoft.com/build-2026/) · 2026-06-02
  • PRODUCT + MODEL CLUSTER (2026-05-14 → 2026-05-21, 6 vendor changelog entries in 8 days -- highest cadence of the year): (1) **2026-05-14 GitHub Copilot app in technical preview** -- standalone Copilot app outside the editor, paralleling the OpenAI Codex mobile / Cursor / Anthropic 'agent goes where you go' pattern. (2) **2026-05-19 Gemini 3.5 Flash GA inside Copilot** -- same day as Google's I/O 2026 reveal (no aggregator delay; vendor-to-vendor handoff). Added to the Pro+/Business/Enterprise model picker. (3) **2026-05-20 Auto model selection routes by task in VS Code** -- Copilot auto-picks among the lineup based on task type rather than user-pinned default. (4) **2026-05-20 Semantic issue search in Copilot Chat** -- chat queries the issues backlog with semantic understanding (find related, dedupe, surface stale). (5) **2026-05-20 Updates to available models in Copilot on web** -- model-picker parity refresh across surfaces. (6) **2026-05-21 Issue fields public preview** -- organization-wide rollout of issue-field customization. Net effect: 6 customer-facing changelogs in the 8-day window ahead of the 6/1 billing cliff.Source: GitHub Changelog (github.blog/changelog -- 5/14 copilot-app-tech-preview, 5/19 gemini-3-5-flash-ga, 5/20 auto-model-selection-vscode, 5/20 semantic-issue-search-copilot-chat, 5/20 updates-to-available-models-on-web, 5/21 issue-fields-public-preview) · 2026-05-21
  • MODEL LINEUP REFRESH (2026-05-17 + 2026-05-18): (1) **GPT-5.3-Codex is now the base model for Copilot Business and Enterprise** (vendor-primary github.blog 2026-05-17), replacing GPT-4.1 ahead of its 6/1 retirement. 1x multiplier, included in standard premium-request quotas; meaningfully faster base-tier completions + chat. (2) **Copilot Cloud Agent added two low-multiplier models 2026-05-18**: Claude Haiku 4.5 and GPT-5.4-mini, both at 0.33x multiplier -- positioned as the 'fast, cost-efficient' default for simple cloud agent tasks. Default cloud-agent runs now route to these tiers; users can still override to Sonnet 4.6 / Opus 4.7 / GPT-5.5 for complex work, but the new default cuts per-task spend ~3x. Material change for any team using Copilot agent mode on PR cleanup / test scaffolding -- audit your spend dashboard after 5/18Source: GitHub Changelog (github.blog/changelog/2026-05-17-gpt-5-3-codex-is-now-the-base-model-for-copilot-business-and-enterprise, github.blog/changelog/2026-05-18-copilot-cloud-agent-fast-cost-efficient-models-for-simple-tasks) · 2026-05-18
  • MODEL RETIREMENT WENT LIVE 2026-05-15: **Grok Code Fast 1 fully removed** from Copilot Chat, inline edits, ask + agent modes, and code completions (the 5/8-announced T-7 cliff landed on schedule, accelerated by xAI's own deprecation that day -- see grok.ts 8-model retirement). Anyone with grok-code-fast-1 pinned in a team config saw failures starting 5/15; replacements are GPT-5 mini or Claude Haiku 4.5. **GPT-4.1 retirement 2026-06-01 still on schedule** (T-13 from today). Joins the GPT-5.2 + GPT-5.2-Codex 6/1 batch already documented, making 6/1 a triple-deprecation cliffSource: GitHub Changelog (github.blog/changelog/2026-05-08-upcoming-deprecation-of-grok-code-fast-1, github.blog/changelog/2026-05-08-upcoming-deprecation-of-gpt-4-1), xAI 5/15 retirement docs · 2026-05-15
  • TWO MODEL DEPRECATIONS ANNOUNCED 2026-05-08 (Grok Code Fast 1 was T-7, now PAST-TENSE per 5/15 entry above): (1) **Grok Code Fast 1 retired 2026-05-15** -- accelerated 'due to the model provider deprecation' on xAI's side. Affected Copilot Chat, inline edits, ask + agent modes, code completions. Replacement: GPT-5 mini or Claude Haiku 4.5. (2) **GPT-4.1 retires 2026-06-01** -- joins the GPT-5.2 + GPT-5.2-Codex 6/1 batch. Replacement: GPT-5.5. Adds to the broader 6/1 usage-based-billing transition cliff: by June 1, model lineup loses GPT-5.2, GPT-5.2-Codex, GPT-4.1, AND moves to AI Credits metering simultaneouslySource: GitHub Changelog (github.blog/changelog/2026-05-08-upcoming-deprecation-of-grok-code-fast-1, github.blog/changelog/2026-05-08-upcoming-deprecation-of-gpt-4-1) · 2026-05-08
  • ORIGINAL 5/8 ANNOUNCEMENT (now past-tense, see 5/15 + 5/08 entries above): (1) Grok Code Fast 1 retires 2026-05-15 -- only T-7 from announcement, accelerated 'due to the model provider deprecation' on xAI's side. Affects Copilot Chat, inline edits, ask + agent modes, code completions. Replacement: GPT-5 mini or Claude Haiku 4.5. If you have Grok Code Fast 1 pinned in any team config, it stops working in a week. (2) **GPT-4.1 retires 2026-06-01** -- joins the GPT-5.2 + GPT-5.2-Codex 6/1 batch already documented. Replacement: GPT-5.5. Adds to the broader 6/1 usage-based-billing transition cliff: by June 1, model lineup loses GPT-5.2, GPT-5.2-Codex, GPT-4.1, AND moves to AI Credits metering simultaneouslySource: GitHub Changelog (github.blog/changelog/2026-05-08-upcoming-deprecation-of-grok-code-fast-1, github.blog/changelog/2026-05-08-upcoming-deprecation-of-gpt-4-1) · 2026-05-08
  • MODEL LINEUP SHRINK (2026-05-07): GitHub Copilot retires Claude Sonnet 4 from its model lineup, mirroring the upstream Anthropic API retirement that took effect 2026-06-15 (now confirmed retired on Anthropic's deprecations page). If your Copilot org/team had pinned Sonnet 4, it is now unselectable -- migrate to Sonnet 4.6 (1x multiplier, same as Sonnet 4) or Opus 4.7 (15x multiplier, dramatically higher cost). Adds to the 2026-06-01 GPT-5.2 + GPT-5.2-Codex deprecation already documentedSource: GitHub Changelog (github.blog/changelog), Anthropic deprecations page · 2026-05-07
  • MULTI-CHANGE BATCH heading into 2026-06-01 billing cutover (vendor-primary verified 2026-05-04): (1) GPT-5.2 + GPT-5.2-Codex DEPRECATION announced 2026-05-01 -- both retired across all Copilot experiences (Chat, inline edits, ask + agent modes, code completions) on 2026-06-01, except GPT-5.2-Codex remains in Copilot Code Review temporarily. Migrate to GPT-5.5 (7.5x multiplier) or GPT-5.4 (1x). (2) Business self-serve signup PAUSED 2026-04-22 -- joining Pro/Pro+/Student paused since 4/20. New Business orgs get Free or Team only; existing Business customers unaffected. (3) Copilot Code Review begins consuming GitHub Actions minutes on 2026-06-01 alongside the existing premium-request charge -- code reviews now bill TWICE (review request + Actions runtime). Budget impact for teams running review on every PR. (4) Verified multiplier table (paid plans, docs.github.com/en/copilot/concepts/billing/copilot-requests as of 2026-05-04): Claude Opus 4.7 = 15x; Claude Opus 4.6 = 3x (fast-mode preview = 30x); Sonnet 4/4.5/4.6 = 1x; Haiku 4.5 = 0.33x; Gemini 3.1 Pro = 1x; Gemini 3 Flash = 0.33x; GPT-5.5 = 7.5x; GPT-5.4 = 1x; GPT-5.4 mini = 0.33x; GPT-5.4 nano = 0.25x; GPT-5.3-Codex = 1x. Aggregator-circulated 'Opus 4.7 = 27x' and 'Gemini 3.1 Pro = 6x' figures are WRONG vs vendor primarySource: GitHub Blog (github.blog/changelog/2026-05-01-upcoming-deprecation-of-gpt-5-2-and-gpt-5-2-codex), (github.blog/changelog/2026-04-22-pausing-new-self-serve-signups-for-github-copilot-business), (github.blog/changelog/2026-04-27-github-copilot-code-review-will-start-consuming-github-actions-minutes-on-june-1-2026), docs.github.com Copilot multiplier table · 2026-05-04
  • PRICING REFORM ANNOUNCED 2026-04-27 (effective 2026-06-01): All Copilot paid plans transition off the 'premium requests' metering system to a usage-based AI Credits model. Each tier's monthly price now ALSO becomes its monthly credit allotment ($10 Pro = $10 in credits; $39 Pro+ = $39 in credits; $19 Business = $19/user; $39 Enterprise = $39/user). Above the allotment, requests pay-as-you-go at standard token API rates (input + output + cached). Code completions and Next Edit suggestions remain free across all plans and DO NOT consume credits. Annual subscribers stay on the existing premium-request model until their current plan expires, then drop to Free with a one-time prorated credit transfer covering remaining value. This is a direct response to agent-workload economics that VP Joe Binder cited as the reason for the 4/20 signup pauseSource: GitHub Blog: 'GitHub Copilot is moving to usage-based billing' (github.blog/news-insights/company-news/github-copilot-is-moving-to-usage-based-billing/) · 2026-04-27
  • CRITICAL (2026-04-20): GitHub paused NEW signups for Copilot Pro, Pro+, and Student plans. Copilot Free remains open. Existing subscribers keep access but face tightened usage caps -- VP Joe Binder cited agent workloads consuming more compute than subscription economics support. Refund window for dissatisfied subscribers: 2026-04-20 through 2026-05-20 via GitHub Support. If you were planning to buy Pro or Pro+, you currently cannot. Cursor, Claude Code, and Windsurf are the obvious alternatives during the pauseSource: GitHub Changelog (github.blog/changelog/2026-04-20), The Register, AfterDawn · 2026-04
  • Claude Opus models were pulled from the Pro tier on 2026-04-20. Opus 4.7 remains on Pro+ only; Opus 4.5 and 4.6 are being removed from Pro+ as well. If you specifically subscribed to Pro for Opus access, your tier has materially changed -- the $10 Pro plan now routes to GPT-5.4 + Gemini 3.1 + smaller open models by defaultSource: GitHub Changelog 2026-04-20, AfterDawn · 2026-04
  • EFFECTIVE 2026-04-24 (yesterday): GitHub now trains its AI models on Copilot Free/Pro/Pro+ interaction data by default -- inputs, outputs, code snippets, file names, navigation patterns, and chat comments. Business + Enterprise plans remain excluded. To opt out: settings > Copilot > Features > 'Allow GitHub to use my data for AI model training' (toggle off). Users who previously opted out of product-improvement data collection have that preference preserved. Big differentiator vs. Cursor and Claude Code, both of which do NOT train on user code by defaultSource: GitHub blog (github.blog/news-insights/company-news/updates-to-github-copilot-interaction-data-usage-policy/), The Register · 2026-04-24
  • New 'premium requests' currency system rolled out in early 2026 -- Pro plan now gets 300 premium requests/mo, Business 1,500, Enterprise 1,000. Beyond those quotas, requests pay-as-you-go. Check the new plan details page before assuming legacy 'unlimited chat' behavior still appliesSource: GitHub Copilot plans page · 2026-03
  • Agent mode reached GA on VS Code and JetBrains in March 2026 -- Copilot can now run multi-step agentic workflows rather than just reply to chat messages. Still behind Cursor Composer 2 and Claude Code on autonomous multi-file work but closes the gap meaningfullySource: GitHub Copilot blog · 2026-03
  • Copilot chat occasionally loses context mid-conversation in larger projects, giving irrelevant suggestionsSource: GitHub Community Discussions · 2026-03

Best for

Teams already on GitHub who want the broadest frontier-model picker in any coding tool (GPT-6 Astra, Sol and Luna, Claude Opus 5.5, Sonnet 5.5 and Fable 5.1, Gemini 3.8 Flash, Grok 4.7 as of 2026-09-28) with enterprise-managed defaults, agent-operation guardrails and usage analytics. Business/Enterprise signups reopened 2026-09-03; Free-tier users still get best-in-class inline completions.

Not for

Teams that pin a single model and expect it to stay: the 10/02 slate removed four models (Gemini 3.5 and 3.6 Flash, Kimi K2.7 Code, Claude Opus 4.7) and the 10/19 slate removed six more, and since 10/22 (rolled out through 10/26) a global default policy switches on new GA features for Business and Enterprise unless an admin configures otherwise. Also not for anyone who wants a fixed monthly bill -- every new frontier model (GPT-6.1 Sol, Opus 5.5, Sonnet 5.5, GPT-6 Sol/Luna, Grok 4.7) is metered at provider list price above the included AI Credits. Individual buyers: the plans page now shows Get started on Pro, Pro+ and Max, but GitHub has not announced the end of the April signup pause, so confirm before you budget around it.

Our Verdict

GitHub Copilot spent the spring of 2026 retrenching -- signups paused on 2026-04-20, Opus pulled from Pro, caps tightened -- and the summer and early autumn rebuilding around usage-based billing (AI Credits from 6/01), enterprise-managed model policy (GA 8/26) and, in the week of 2026-09-22, the fastest model intake we have recorded: Claude Opus 5.5, GPT-6 Sol and GPT-6 Luna on the day their vendors shipped them, Claude Sonnet 5.5 six days later, all auto-enabled. That makes Copilot the only coding tool where the newest Anthropic, OpenAI and xAI models are selectable side by side on launch day, and the agent platform underneath it (local sandboxing, OpenTelemetry export, assisted approvals, code-review defaults on every plan) is now credible enterprise infrastructure rather than autocomplete with a chat box. The costs are churn and spend: ten models left the picker in October, a global default-on policy for new features rolled out 10/22 to 10/26, and every frontier model is billed at provider list price past the credits. The plans page listing Pro, Pro+ and Max as purchasable again is the best individual-buyer news since April, pending an actual announcement. At Universe (10/28) GitHub made Copilot Fusion GA, which picks and combines models for you at three effort levels, and began including Copilot in GitHub Enterprise licences for account-team customers. Scores unchanged; the picker is the product.

Sources

  • GitHub Changelog: Usage metrics API adds code review details and addressed rate (2026-11-05) (accessed 2026-11-05)
  • GitHub Changelog: Usage metrics API flags telemetry and version gaps (2026-11-05) (accessed 2026-11-05)
  • GitHub Changelog: GitHub Copilot is now available in Xcode's coding assistant (2026-11-04) (accessed 2026-11-05)
  • GitHub Changelog: GitHub Copilot is now available in JetBrains Air (2026-11-03) (accessed 2026-11-05)
  • GitHub Changelog: BYOK and enterprise custom models in the Copilot usage API (2026-11-03) (accessed 2026-11-05)
  • GitHub Changelog: GitHub intelligence -- Copilot Memory GA, Copilot dreaming preview (2026-10-28) (accessed 2026-11-05)
  • GitHub Changelog: Repository plugin defaults and enterprise plugin controls generally available (2026-10-28) (accessed 2026-11-02)
  • GitHub Changelog: Route and orchestrate models with GitHub Copilot Fusion (2026-10-28) (accessed 2026-10-29)
  • GitHub Changelog: GitHub Enterprise now includes access to GitHub Copilot; Copilot Standalone $10 per user per month (2026-10-28) (accessed 2026-10-29)
  • GitHub Changelog: Generate and edit images directly in Copilot, GPT Image 2.5 Flare (2026-10-28) (accessed 2026-10-29)
  • GitHub Changelog: Copilot code review Max and approvals (2026-10-28) (accessed 2026-10-29)
  • GitHub Changelog: GitHub Copilot app, Universe releases (2026-10-28) (accessed 2026-10-29)
  • GitHub Changelog: Work from anywhere with a new, connected Copilot cloud experience (2026-10-28) (accessed 2026-10-29)
  • GitHub Changelog: Connect your tools to Copilot without all the setup, 26 connectors (2026-10-28) (accessed 2026-10-29)
  • GitHub Changelog: GPT-6.1 Terra in GitHub Copilot (2026-10-26) (accessed 2026-10-29)
  • GitHub Changelog: GitHub Copilot models deprecated on October 19, 2026 (executed; alternatives GPT-6.1 Sol, GPT-6 Luna, Grok 4.7, Gemini 3.8 Flash) (accessed 2026-10-29)
  • GitHub Changelog: Global default for Copilot features generally available, rolled out through Oct 26 (2026-10-22) (accessed 2026-10-29)
  • GitHub Changelog: Multi-repository support for GitHub in Slack and Teams (2026-10-15) (accessed 2026-10-17)
  • GitHub Changelog: Gemini 4 Argon in GitHub Copilot (2026-10-13) (accessed 2026-10-15)
  • GitHub Changelog: Upcoming deprecation of selected GitHub Copilot models in mid-November (2026-10-15) -- Opus 4.8, Opus 5, GPT-5.6 Terra on 2026-11-16 (accessed 2026-10-15)
  • GitHub Changelog: HydraFusion updates and new Max option (2026-10-14) (accessed 2026-10-15)
  • GitHub Changelog: Agentic autofix for code scanning alerts general availability (2026-10-15) (accessed 2026-10-15)
  • GitHub Changelog: AI Scan for pull requests generally available (2026-10-15) (accessed 2026-10-15)
  • GitHub Changelog: Claude Haiku 5.5 in GitHub Copilot (2026-10-07) (accessed 2026-10-12)
  • GitHub Changelog: Local sandboxing for GitHub Copilot now generally available (2026-10-07) (accessed 2026-10-12)
  • GitHub Changelog: Discover local models in GitHub Copilot CLI (2026-10-07) (accessed 2026-10-12)
  • GitHub Changelog: Purpose-built model for leaked secret detection (2026-10-07) (accessed 2026-10-12)
  • GitHub Changelog: Copilot code review -- new organization billing options and controls (2026-10-08) (accessed 2026-10-12)
  • GitHub Changelog: GitHub Copilot weekly releases -- October 5 (2026-10-09) (accessed 2026-10-12)
  • GitHub Changelog: New controls and chat improvements in Copilot for JetBrains (2026-10-10) (accessed 2026-10-12)
  • GitHub Changelog: Update your IDE to restore agent activity in Copilot usage metrics (2026-10-06) (accessed 2026-10-12)
  • GitHub Changelog: Client telemetry user attribution and configurable capture (2026-10-12) (accessed 2026-10-12)
  • GitHub Changelog: Selected models in GitHub Copilot deprecated (2026-10-02) -- Gemini 3.5/3.6 Flash, Kimi K2.7 Code, Claude Opus 4.7 removed; Opus 5.5 named as replacement (accessed 2026-10-02)
  • GitHub Changelog: Copilot code review API support and new default effort level (2026-10-02) (accessed 2026-10-02)
  • GitHub Changelog: GPT-6.1 Sol in GitHub Copilot (2026-09-29) -- Pro+/Max/Business/Enterprise, provider list pricing (accessed 2026-10-02)
  • GitHub Changelog: HydraFusion in VS Code and the GitHub Copilot app (2026-09-30) -- Single / Cascade / Critique workflows (accessed 2026-10-02)
  • GitHub Changelog: GitHub Copilot can now interact with desktop apps with computer use (2026-10-01, public preview) (accessed 2026-10-02)
  • GitHub Changelog: Dynamic workflows in Copilot CLI and the Copilot app (2026-10-01) (accessed 2026-10-02)
  • GitHub Changelog: GitHub Copilot in VS Code, September 2026 releases (2026-10-01, v1.136-1.140) (accessed 2026-10-02)
  • GitHub Changelog: Claude Sonnet 5.5 in GitHub Copilot (2026-09-28) -- Pro/Pro+/Max/Business/Enterprise, GA, list pricing (accessed 2026-09-28)
  • GitHub Changelog: Claude Opus 5.5 is now available in GitHub Copilot (2026-09-22) -- Pro+/Max/Business/Enterprise, text watermark (accessed 2026-09-28)
  • GitHub Changelog: OpenAI's GPT-6 Sol and GPT-6 Luna now available (2026-09-22) -- Sol Pro+ up, Luna Pro up (accessed 2026-09-28)
  • GitHub Changelog: Default Enablement of Copilot features for Copilot Business and Enterprise (2026-09-24) -- policy takes effect October 22 (accessed 2026-09-28)
  • GitHub Changelog: GitHub Copilot weekly releases -- September 21 (2026-09-25) -- model tier map, sandboxing, OTel, Slack/Teams, JetBrains, VS Code 1.139 (accessed 2026-09-28)
  • GitHub Changelog: More ways to request and configure Copilot code reviews (2026-09-23) -- personal settings on every plan, enterprise default effort (accessed 2026-09-28)
  • GitHub Changelog: Local sandboxing in the GitHub Copilot app (2026-09-23, public preview) (accessed 2026-09-28)
  • GitHub Changelog: OpenTelemetry in the GitHub Copilot app (2026-09-22) (accessed 2026-09-28)
  • GitHub Changelog: New features and improvements in Copilot for JetBrains 1.18.0 (2026-09-22) -- assisted approvals, re-edit, plan mode for Codex agent (accessed 2026-09-28)
  • GitHub Changelog: Faster C++ code intelligence with whole codebase indexing (2026-09-22) (accessed 2026-09-28)
  • GitHub Changelog: Updates to GitHub Copilot for Slack and Microsoft Teams (2026-09-25) (accessed 2026-09-28)
  • GitHub Changelog: Agentic autofix now uses Copilot Memory (2026-09-25) (accessed 2026-09-28)
  • GitHub Changelog: Usage metrics API adds pull request review stages (2026-09-25) (accessed 2026-09-28)
  • GitHub: Copilot plans page -- Free, Pro $10, Pro+ $39, Max $100 with Get started buttons, no pause notice (checked 2026-09-28) (accessed 2026-09-28)
  • GitHub Changelog: Grok 4.7 is now available in GitHub Copilot (2026-09-21) -- Pro/Pro+/Max/Business/Enterprise, auto-enabled under global model policy, gradual rollout (accessed 2026-09-21)
  • GitHub Changelog: Upcoming deprecation of selected Copilot models in mid-October (2026-09-18) -- Gemini 3.7 Flash, GPT-5.5, GPT-5.4, GPT-5.4 mini, GPT-5 mini, Grok 4.5 removed 2026-10-19 (accessed 2026-09-21)
  • GitHub Changelog: Copilot code review, an improved review experience (2026-09-18) -- grouped overview, comment titles, resolution reasons, batch commit messages (accessed 2026-09-21)
  • GitHub Changelog: GitHub Copilot weekly releases, September 14 (published 2026-09-18) -- VS Code 1.138 Dev Containers, Sentry canvas, tier rollout (accessed 2026-09-21)
  • GitHub Changelog: Copilot impact dashboard now shows feature engagement (2026-09-17) (accessed 2026-09-21)
  • GitHub Changelog: Agentic CLI customizations now in the usage metrics API (2026-09-17) (accessed 2026-09-21)
  • GitHub Changelog: Copilot budget increase requests are generally available (2026-09-16) (accessed 2026-09-17)
  • GitHub Changelog: GitHub Copilot suggests custom properties definitions (2026-09-15) (accessed 2026-09-17)
  • GitHub Changelog: MAI-Code-1-Flash deprecated -- retired across all Copilot experiences, alternative MAI-Code-1.1-Flash (2026-09-10) (accessed 2026-09-14)
  • GitHub Changelog: Configure cost and quality in Copilot auto model selection -- efficiency/balance/intelligence tiers, 10% auto discount retained (2026-09-14) (accessed 2026-09-14)
  • GitHub Changelog: Enterprise managed permissions for GitHub Copilot agent operations -- block / require approval / allow for shell, files, network (2026-09-09) (accessed 2026-09-14)
  • GitHub Changelog: Enterprise-managed sandbox in Copilot for JetBrains + /ide CLI binding (2026-09-08) (accessed 2026-09-14)
  • GitHub Changelog: Auto-resolution and analysis updates in Copilot code review -- shell tools, ensemble Lite reviews (2026-09-11) (accessed 2026-09-14)
  • GitHub Changelog: GitHub Copilot weekly releases -- September 7: Jira in Copilot app, HydraFusion in CLI, VS Code 1.137 automations + voice mode (2026-09-10) (accessed 2026-09-14)
  • GitHub Changelog: Add VS Code Agents to Copilot usage metrics (2026-09-11) (accessed 2026-09-14)
  • GitHub Changelog: Remediate code quality findings with agentic autofix -- bulk assign up to 25 findings to Copilot (2026-09-09) (accessed 2026-09-14)
  • GitHub Changelog: GPT-6 Astra generally available in GitHub Copilot -- auto-enabled under default model enablement (2026-09-04) (accessed 2026-09-05)
  • GitHub Changelog: Gemini 3.8 Flash in Copilot -- introductory provider pricing through 2026-12-31 (2026-09-03) (accessed 2026-09-05)
  • GitHub Changelog: Claude Fable 5.1 in Copilot -- OFF by default, requires data retention for Anthropic safety classifiers (2026-09-01) (accessed 2026-09-05)
  • GitHub Changelog: Upcoming deprecation of selected Copilot models on 2026-10-02 -- Gemini 3.5/3.6 Flash, Kimi K2.7 Code, Claude Opus 4.7 (2026-09-03) (accessed 2026-09-05)
  • GitHub Changelog: Reopening Copilot Business and Enterprise signups -- upfront seat charging from 2026-10-01 for existing card/PayPal customers (2026-09-03) (accessed 2026-09-05)
  • GitHub Changelog: Enterprise-managed settings support any default model, incl. per-team (2026-09-02) (accessed 2026-09-05)
  • GitHub Changelog: Copilot code review can now approve pull requests (2026-09-01) (accessed 2026-09-05)
  • GitHub Changelog: Content exclusions generally available in Copilot app and CLI (2026-09-02) (accessed 2026-09-05)
  • GitHub Changelog: Copilot code review -- resolution reasons and expanded capabilities; 300-file/20k-LOC cap removed, bot-authored PRs reviewed (2026-08-27) (accessed 2026-08-31)
  • GitHub Changelog: Shared agentic work with Copilot in Microsoft Teams (2026-08-21) (accessed 2026-08-28)
  • GitHub Changelog: The new GitHub Copilot experience in Slack (2026-08-21) (accessed 2026-08-28)
  • GitHub Changelog: Copilot harness GA in Copilot for JetBrains (2026-08-24) (accessed 2026-08-28)
  • GitHub Changelog: Global model policy generally available (2026-08-26) (accessed 2026-08-28)
  • GitHub Changelog: Upcoming changes to Copilot policies and billing (2026-08-28) (accessed 2026-08-28)
  • GitHub Changelog: Enterprise managed settings in GitHub Copilot for JetBrains -- MCP allowlist, Autopilot kill switch (2026-08-18) (accessed 2026-08-20)
  • GitHub changelog: MAI-Code-1.1-Flash available in GitHub Copilot -- 0.25x multiplier, vision, off by default for orgs (2026-08-11) (accessed 2026-08-13)
  • GitHub changelog: Upcoming deprecation of MAI-Code-1-Flash -- retires 2026-09-10 (2026-08-11) (accessed 2026-08-13)
  • GitHub changelog: Agent Plugins 1.0 in VS Code, Copilot CLI and the Copilot app -- GA all plans (2026-08-12) (accessed 2026-08-13)
  • GitHub changelog: Copilot memory and Ollama BYOK in Copilot for JetBrains (2026-08-11) (accessed 2026-08-13)
  • GitHub changelog: Per-model token breakdown in the usage report -- Business/Enterprise/Individuals (2026-08-11) (accessed 2026-08-13)
  • GitHub changelog: Copilot on web expands conversation controls -- token spend indicators, all plans (2026-08-10) (accessed 2026-08-10)
  • GitHub changelog: Copilot code review effort levels are generally available -- Lite/Balanced, Pro through Enterprise (2026-08-07) (accessed 2026-08-10)
  • GitHub changelog: Copilot weekly releases August 3 -- Auto shows model + AI credit/cache details (2026-08-07) (accessed 2026-08-10)
  • GitHub changelog: Kimi K3 is now available in GitHub Copilot (2026-08-06) (accessed 2026-08-06)
  • GitHub changelog: Upcoming deprecation of GitHub Spark on github.com (2026-08-04) (accessed 2026-08-04)
  • GitHub changelog: Customize the reasoning level for Copilot cloud agent (2026-08-03) (accessed 2026-08-04)
  • GitHub changelog: Retiring the Copilot billing preview app (2026-08-04) (accessed 2026-08-04)
  • GitHub changelog: Grok 4.6 is now available in GitHub Copilot (2026-08-14) -- all paid SKUs, policy off by default for orgs (accessed 2026-08-17)
  • GitHub changelog: Gemini 3.7 Flash is now available in GitHub Copilot (2026-08-13) -- all paid SKUs, admins must enable the Preview policy (accessed 2026-08-17)
  • SpaceXAI: Grok 4.6 in GitHub Copilot (2026-08-14) -- $2/$6 per 1M direct from the console, unchanged from Grok 4.5 (accessed 2026-08-17)
  • GitHub changelog: Grok 4.5 is now available in GitHub Copilot (2026-07-28) (accessed 2026-07-29)
  • GitHub changelog: Claude Opus 5 is now available in GitHub Copilot (2026-07-24) (accessed 2026-07-29)
  • GitHub Changelog: Claude Sonnet 5 GA for GitHub Copilot (2026-06-30) (accessed 2026-07-04)
  • GitHub Changelog: GitHub Models fully retired 2026-07-30 (announced 2026-07-01) (accessed 2026-07-04)
  • GitHub Changelog: Updates to GitHub Copilot billing and plans -- usage-based billing live (2026-06-01) (accessed 2026-06-02)
  • Microsoft AI: Introducing MAI-Code-1-Flash (2026-06-02) (accessed 2026-06-02)
  • GitHub Changelog: GPT-5.2 + GPT-5.2-Codex deprecation (2026-05-01, retiring 2026-06-01) (accessed 2026-05-04)
  • GitHub Changelog: Business self-serve signup pause (2026-04-22) (accessed 2026-05-04)
  • GitHub Changelog: Code Review Actions minutes billing (2026-04-27, effective 2026-06-01) (accessed 2026-05-04)
  • GitHub docs: Copilot requests + multiplier table (accessed 2026-05-04)
  • GitHub Blog: Copilot moving to usage-based billing (2026-04-27, effective 2026-06-01) (accessed 2026-04-30)
  • GitHub Changelog 2026-04-20: Copilot plan changes for individuals (accessed 2026-04-21)
  • The Register: GitHub grounds Copilot account creation (accessed 2026-04-21)
  • AfterDawn: Usage limits + Opus removal (accessed 2026-04-21)
  • GitHub Copilot plans (accessed 2026-04-21)
  • GitHub blog: Copilot interaction data usage policy update (effective 2026-04-24) (accessed 2026-04-25)
  • GitHub: Copilot Pro/Business training policy (accessed 2026-04-21)

Explore more GitHub Copilot rankings

Deeper leaderboards, benchmarks, task-specific tier lists, and status/pricing pages for GitHub Copilot.

Full AI Code Assistants tier list
Where GitHub Copilot ranks vs every competitor in its category
Best AI tools to debug code
Coding assistants that read a stack trace or failing test and propose a fix with a reasoned explanation.
Best AI tools to explain code
Tools that walk through what a function, file, or repo actually does in plain English.
Best AI tools to write unit tests
Tools that generate unit tests, including edge cases and mocks, from existing source code.
Is GitHub Copilot down?
Outage check plus rolling log of known issues
GitHub Copilot pricing
Every tier and what's included
GitHub Copilot alternatives
Comparable tools at every tier

The Tier List Tuesday

Weekly newsletter: tier movers, new entrants, and the VS of the week. Built from our daily AI-tool sweeps. No spam, unsubscribe anytime.

Alternatives to GitHub Copilot

Cursor logo

Cursor

**Security incident (2026-10-29): a malicious copy of Anthropic's Claude Code extension, versions 2.1.312 and 2.1.313 for Windows and macOS, was served from Cursor's own VS Code Extension Marketplace from about 2:27am to 1:35pm PT and tried to plant a persistent SSH backdoor; Cursor pulled it, shut the marketplace down, rotated keys, rebuilt it and published a check script, remediation scripts and IOCs.** **Nov 5: Cursor now builds and hosts live apps -- a cloud agent scaffolds the web app, sets up Postgres if needed, adds sign-in through Cursor SSO and provisions hosting, runs security checks before anything is published, and shares the app by link (publicly on individual plans). It is on all paid plans with cloud agents enabled.** **Cursor released the Agent SDK as generally available on 2026-10-20: the framework behind Bugbot, Security Reviewer and Rollouts, for defining agents as code, testing them with evals and deploying them to Cursor or your own infrastructure with one command.** **Oct 16: cloud agents can now build interactive canvases and share them by link, the Environments API entered beta for creating and managing agent environments in code, and cloud agent environments hold up to 300 repositories with large environments loading 5x faster.** **Cursor's own plans closed to new customers on 2026-10-12: Cursor is now sold through The X Subscription (Free, Plus $8, Premium $30, Super $100, Ultra $200) with one usage pool shared with Grok, at least twice the coding usage on SpaceXAI's models -- but Claude, GPT and Gemini are no longer in the allowance, billed at API list price on Premium and up and unavailable on Plus. On 2026-10-06 the Cursor iOS app gained remote control of local agents.** **Cursor launched two bots for the last mile of shipping code on 2026-09-23 (Teams and Enterprise): Rollouts, which attaches a monitor to every pull request and watches the change as it deploys per environment -- verified healthy, regression detected, or inconclusive -- and Security Review, which posts one comment per PR reporting exploitable bugs with severity, attack path and a proposed fix.** Rollouts is the Cursor rebuild of Firetiger's Change Monitors on the Bot Development Kit; it can open a revert PR or hand a regression to a cloud agent but does not merge or roll back on its own. Models: Anthropic's 9/28 Sonnet 5.5 post quotes SpaceXAI's Sualeh Asif on CursorBench 4.0 -- Sonnet 5.5 at 55.5% 'second only to Opus 5.5' (57.8%); Grok 4.7 and its Cursor-only Fast variant arrived 9/21. **Cursor Projects (9/10, beta)** and the 11/12 OpenAI shutoff watch continue.

A
8.3/10
Free tierFrom $0
Cursor 3's agent-first redesign (April 2...Composer 2 is Cursor's own frontier codi...
Updated 2026-11-05
Devin Desktop (formerly Windsurf) logo

Devin Desktop (formerly Windsurf)

Windsurf is now **Devin Desktop** -- Cognition retired the Windsurf brand via OTA update on June 2, 2026. Same editor, plans, pricing, settings, and extensions; the bundled agent is now 'Devin Local' and Devin Cloud agent access starts on the $20 Pro plan. Agent Command Center, Spaces, and Devin Review all carry over

B
7.5/10
Free tierFrom $0
Windsurf 2.0 (launched 2026-04-15) is a ...Embedded Devin cloud agent (via Cognitio...
Updated 2026-07-09
Tabnine logo

Tabnine

**Tabnine was acquired by Tricentis on 2026-07-30 and is being folded into the Tricentis Agentic Quality Engineering Platform as its Enterprise Context Engine; tabnine.com now redirects to tricentis.com and the public price list is gone, though the VS Code extension (v3.335.0, about 9.6M installs) and the docs are still live.** Privacy-first AI code completion with on-premise deployment -- now a testing-company asset rather than a standalone Copilot rival.

C
6.3/10
Free tierFrom $0
Privacy-first approach -- code never lea...Works as a plugin in any major IDE (VS C...
Updated 2026-10-06
Claude Code logo

Claude Code

**Claude Code 2.1.317 and 2.1.318 (2026-11-02 and 2026-11-03): claude auth status and claude -p JSON now report when a stored claude.ai sign-in expires, with a stderr warning from ten days out; the Claude apps gateway gets a cost report endpoint; Claude asks before a tool call sends your account email to the network; and Claude Code no longer installs the IDE extension on its own. Message threads are now on by default for most Claude Code users, cutting median time to first token by about 35% on 500K-1M-token requests.** **Claude Code 2.1.314 to 2.1.316 (2026-10-30 to 2026-10-31): a setting that blocks writes outside the working directories in every permission mode, idle compaction for long conversations in VS Code and Claude Desktop, pooled group budgets on the Claude apps gateway, and notice that fast mode on Opus 4.8 is being retired, with removal on 2026-11-30. Separately, on 2026-10-29 a malicious copy of the Claude Code extension (labelled 2.1.312 and 2.1.313) was served for about eleven hours from Cursor's own extension marketplace; Cursor reports no evidence the official distribution was affected.** **Claude Code 2.1.303 to 2.1.312 (2026-10-18 to 2026-10-27): 2.1.304 added /auto-mode-setup, Code Review levels (Quick, Standard, Deep) and a warning on Pro and Max before resuming an old conversation would use a large share of the 5-hour limit; 2.1.307 added /claude-api api-version for Claude API version 2026-10-21; 2.1.311 makes ultrathink run a turn at xhigh effort on Opus 5.5, Sonnet 5.5, Haiku 5.5 and Fable 5.1, and has Bedrock, Vertex, Foundry and Mantle load model settings from a signed Anthropic file unless you turn it off.** **2.1.302 (2026-10-16) added CLAUDE_CODE_AUTO_MODE_MODEL to choose the model behind the client-side auto-mode classifier, an Idle compaction on/off row in /config, and a notice in /design-sync and /design-login that standalone Claude Design (claude.ai/design) closes by 2026-12-14; it also fixed file writes on exFAT/FAT32 drives on Windows and stopped prompt suggestions spending credits.** **2.1.299 (2026-10-14) added idle compaction -- a 200K+ token conversation left idle about 54 minutes is compacted before its one-hour prompt cache expires -- plus /effort and /fast defaults in agent view and /usage warnings for max effort and large tool results; 2.1.300 (10/14) brought /advisor to Claude Platform on AWS; 2.1.301 (10/15) lets unattended retry mode give up on far-off usage-limit resets and makes the fable alias fall back to the previous Fable model.** **Claude Haiku 5.5 became the default Haiku in 2.1.293 (2026-10-07) at $0.10 / $0.50 per 1M below 100K tokens; 2.1.295 (10/08) added onFailure block for hooks, OSC 7501 terminal status, per-upstream model lists on the Claude apps gateway and a 16,384-character limit on tool-search MCP descriptions, and 2.1.296 (10/09) added an allow_large Read option, per-subagent autoCompactWindow and a code policy key for Claude Desktop's Code tab.** **Three releases in two days -- 2.1.290 (2026-10-05), 2.1.291 and 2.1.292 (both 10/06): claude attach and claude logs by session name, a /claude-api managed-agents-onboard command, an effort parameter on the Agent tool, --marketplace on claude plugin install, a configurable 529 retry delay, WebFetch continuation past 100,000 characters, local MCP servers defaulting to protocol 2026-07-28, and a long run of security fixes for symlink and UNC-path reads; no model or pricing changes.** Anthropic's terminal-based coding agent that reads your whole repo and makes real changes -- not just suggestions. **Claude Opus 5.5 became the default Opus model in 2.1.280 (2026-09-22) and Claude Sonnet 5.5 the default Sonnet in 2.1.284 (2026-09-28)**: Opus 5.5 at $4/$20 with $0.20 cache reads and fast mode at $8/$40 for up to 2.5x speed, Sonnet 5.5 at an unchanged $2/$10 with far fewer tokens per task. Both are 1M context. Fable 5.1 (2026-09-01) remains the top model and still defaults to High effort in Claude Code specifically. **Four more releases 9/29-10/02 (2.1.285-2.1.288) added Claude Mods (plugins that modify deeper behaviour) with a built-in 'You should know' side agent, `claude --desktop` to hand a session to the desktop app, an `allowedProviders` managed setting, a WebFetch kill switch, `--max-findings` on /code-review, and continuation from partial responses when the API times out mid-turn.** Also since 9/22: `/doctor prompt-audit`, `deniedModels`, dollar amounts on spend limits and `attribution: false`.

B
7.8/10
From $20
Reads and understands your entire codeba...Actually executes code, runs tests, and ...
Updated 2026-11-05
Lovable logo

Lovable

**Lovable's Nov 2-4 changelog: Dreaming, a background pass on paid plans that rewrites or removes stale project memory and AGENTS.md rules without spending credits (Nov 4); project monitoring that now also catches errors in visitors' browsers (Nov 2, paid plans); AGENTS.md and CLAUDE.md files in subfolders now followed (Nov 3); and Enterprise abandonment protection so a project is never auto-deleted for inactivity (Nov 3).** **Lovable's Oct 28 to Nov 1 changelog: memory that carries the preferences, decisions and facts you state into later conversations, per project and per person, readable and editable in settings (Oct 30); importing a project from Replit, Bolt, Base44, v0 or Manus as a ZIP, or an existing Wix, Squarespace or Webflow site by URL, with the first import free (Oct 29); Deep scan security reviews started from the chat (Oct 29); a downloadable SBOM in SPDX 2.3 or CycloneDX 1.5 (Nov 1); and an email before your credits run out (Oct 28).** **Lovable's Oct 16 to 27 changelog: Enterprise publishing policies and an Action center for admins, OIDC sign-in for app users, tests written for changed business rules, ByteDance Seedream 5 Pro and Seedance 2.5 for app AI features, organizations that group several Enterprise workspaces, and Events and Hooks for Business and Enterprise. Gemini 2.5 chat models left app AI features on 10/20, with requests now routed to Gemini 3.8 Flash or Gemini 3.1 Flash Lite.** **Lovable launched Growth tools on 2026-10-14: set a goal such as your first 100 signups and it builds a plan from your app and traffic, prepares SEO, listings and creative, and launches Google, YouTube, Facebook and Instagram ad campaigns you approve, with no extra Lovable fee. Same week: Sign in with Lovable for any website, 60-day restore for deleted projects, Gemini 4 Argon (10/13) and GPT-6.1 Sol (10/15) for app AI features, and Enterprise audit-log export to SIEMs.** **Lovable shipped credit usage limits and alerts for workspaces (10/07), invite-only publishing on Free and Pro (10/08), nine new app connectors plus a Loom chat connector, Drizzle migration files for Cloud database changes, chat-to-chat task handoff and Enterprise workspace publishing domains (10/09).** Describe the app you want in plain English and watch it build itself. **New this week: apps built on Lovable can generate music with Google's Lyria 3 Clip Preview (about 30-second clips) and Lyria 3 Pro Preview (tracks up to 184 seconds) from a prompt or an image (2026-10-05, credits charged per finished track), beehiiv and Ramp connectors arrived, Lovable can restore its own work after a Git push overwrites it, and GitHub secret scanning now detects leaked Lovable API keys (10/05). Earlier: live voice conversations through GPT Live (2026-10-01, TanStack Start apps, uses credits), Enterprise workspaces can pin AI processing to EU providers (10/01, costs more credits), Business and Enterprise API keys can be IP-allowlisted (10/01), Lovable can rotate a Cloud database password from chat (9/30), and the Lovable MCP server works inside Muse by Meta (9/29).** **Chat with Lovable (2026-09-21)** adds a free daily chat allowance for thinking, asking and planning without changing code, plus voice conversations and a Telegram surface; apps built on Lovable can now call **Claude Opus 5.5 (9/23)** and **GPT-6 Sol and Luna (9/25)** for their own AI features, and the **Lovable API (9/17)** automates workspaces. $500M ARR (June 2026) and ~1M new projects a week

B
7.8/10
Free tierFrom $0
Growth tools (2026-10-14) carry a Lovabl...The ease of use is unmatched -- describe...
Updated 2026-11-05
Devin logo

Devin

**Nov 4: SWE-2 Flash, a 276B / 12B-active coding model Cognition post-trained from Thinking Machines' Inkling-Small, is in Devin Desktop and CLI at one-fifth the price and three times the speed of SWE-2 and free through Nov 30; the weights are held back for a vetted-partner programme. Nov 5: Devin Review 2.0 makes Devin Review the default PR reviewer, at a typical $0.25-$0.75 per PR after an 82% cost cut since August.** **Nov 2: hierarchical budgets let enterprise admins split one account-wide ACU limit into department, project and per-user budgets up to three levels deep, with delegated owners who can grant temporary bumps inside their own subtree.** **Corridor, which calls itself the security layer for AI-native teams, joined Cognition on 2026-10-28 to work on Devin Security; the announcement gives no deal terms.** **Devin Security 2.0 (2026-10-26) runs deep scans on Claude Mythos and other frontier cyber models for enterprises, ranks findings against a threat model built with your team and fixes them in parallel sessions, and Cognition is putting hundreds of millions of dollars in Devin credits toward free scans and remediation. Devin Oncall opened to all organizations on 10/19, and Playbooks are becoming Skills (new accounts from 10/24 start with Skills). Current plans: Free, Pro $20, Max $200, Teams $80 plus $40 per dev seat, Enterprise.** **Oct 16 release notes: /devin review and /devin prompts now work on GitLab merge requests, Devin Review shows Bitbucket Cloud merge readiness, Devin gained built-in git tools for issues, labels, assignees and branch protection plus PR search across GitHub, GitLab, Azure DevOps, Gerrit and Perforce Swarm, and queued messages stay editable while Devin is idle.** **Devin now has memory: the 2026-10-09 release notes add personal memory across sessions with a nightly 'dreaming' pass that refines what it keeps (admins can force it on or off), cut the session resume window from 30 to 14 days, load a repository's AGENTS.md on setup and open Devin Ultra to zero-data-retention enterprises; 10/07 added a notification inbox. Anthropic's Haiku 5.5 launch lists it as a Devin Fusion sidekick option.** **10/02 and 10/05 release notes: a Terminal tab in the session workspace, keyboard shortcuts and a command palette for everyone, Devin Review effort levels set per trigger with an enterprise default, single-code-scan fetch via API and MCP, Azure DevOps review summaries posted as closed threads, Perforce snapshot builds given three hours, and Microsoft Teams access settings now enforced.** **9/30 release notes: Devin Review now routes findings on a Devin-opened PR back to the session to fix first (only what it cannot fix gets posted), automations gain preflight scripts, and dropped-connection messages are no longer lost.** **Cognition crossed a one billion dollar run rate on 2026-09-25** -- '$1B in annualized revenue run rate', less than two years after Devin's GA, naming GE Aerospace, Rivian, Rohlik and Exa -- and opened a Sao Paulo hub on 9/22 with Itau (over 75% of its technology teams on Devin, 300,000+ repositories documented, .NET-to-Java migrations 6x faster), Nubank (a multi-million-line monolith migration 'from years to weeks' at over 20x lower cost), Santander, Natura and EBANX as customers. **Devin Fusion (9/11)** -- the lead/sidekick two-model harness measured at 23-46% lower cost per task -- and the 9/15 AWS multi-year agreement stand.

B
7.4/10
Free tierFrom $0
Genuine autonomy -- you can describe a t...Desktop / GUI testing via computer-use (...
Updated 2026-11-05
Replit logo

Replit

Cloud IDE with an AI agent that builds full apps from prompts. **Free Mode (2026-08-18)**, powered by OpenAI's GPT-5.6 Luna, lets Core and Pro subscribers chat and run everyday tasks without spending credits -- up to 30 hours of chat a month on Core -- and **Intelligent Model Routing (2026-08-26)** now picks the model per task at what Replit measured as 65% lower cost than the old Max Mode. Replit acquired the charting startup Atta on 2026-09-25. Agent 4 (May 2026) added parallel task execution

B
7.0/10
Free tierFrom $0
Zero setup -- open a browser, describe y...Full development environment in the clou...
Updated 2026-09-28
Codex (OpenAI) logo

Codex (OpenAI)

**Nov 4: OpenAI agreed to acquire Pierre Computer Company, maker of the Trees and Diffs libraries the Codex app already uses to browse files and review changes and of the Code Storage git service, and Images 2.6 (GIF output, faster edits) arrived in Codex on every plan.** **Oct 29: OpenAI doubled the Codex and Work allowance on Pro 100 and Pro 500 and called off the planned Pro 200 cut, so every Pro 200 subscriber now gets the higher allowance; Plus usage is unchanged.** **GPT-6.1 Terra (2026-10-26) is the new cheap workhorse in Codex and the API: gpt-6.1-terra costs $1 input / $0.05 cached / $5 output per 1M tokens, half of GPT-6.1 Sol, and OpenAI reports 50.3% on Terminal-Bench 4.0 at max effort against 21.5% for GPT-5.6 Terra, while saying Sol and GPT-6 Astra remain the better choice for the hardest coding. The Decisions API went GA on 10/22.** **GPT-6.1 Sol Ultrafast is now on the API rate card -- $12 input / $0.60 cached / $15 cache writes / $60 output per 1M up to 272K tokens, six times Standard, closing the DevDay 'coming days' promise -- and on 2026-10-09 the Codex desktop app gained Composer predictions (beta), which suggests your next message after each response for personal Pro users on GPT-6 Astra and GPT-6.1 Sol.** **OpenAI and Atlassian expanded their partnership on 2026-10-06: GPT-6 family models (GPT-6 Astra and the GPT-5.6 series are named) will power agents across Atlassian's platform and Rovo, more than 3,000 Atlassian developers already use Codex, and the Atlassian and Teamwork Graph CLI plugins connect Codex and ChatGPT to Jira, Confluence, Looms and Bitbucket pull requests.** **Two October notes: EU Codex output will carry OpenAI's invisible textGrain text watermark within weeks (API customers can opt in worldwide from 10/05), and the new Enterprise token rate card prices Codex at API Standard rates, names GPT-5.3-Codex as the code-review model and GPT-5.6 Luna as the auto-review model, and puts Astra Ultrafast at 6x.** GPT-6.1 Sol Ultrafast is still not on the API rate card as of 10/05. **DevDay 2026 (2026-09-29) was mostly a Codex release: GPT-6.1 Sol in Codex and the API at $2/$10 per 1M with cached input cut to $0.10, Codex Cloud (reusable cloud environments, tasks that keep running while your laptop sleeps, on Plus and up), a refreshed CLI with voice steering and an /agents view, Code Review in the desktop app with automatic cloud first passes, Codex Security Cloud with Daybreak Blue models and no separate Daybreak application, and the Agents API gaining computer use** -- plus GPT-6 Astra Ultrafast (300 tokens per second in Codex) on Pro 500 and Enterprise, with GPT-6.1 Sol Ultrafast 'coming soon'. GPT-6.1 Sol nearly matches GPT-6 Astra on DeepSWE at one-fifth of the token price and also landed in GitHub Copilot the same day. Still in place: GPT-6 Sol and Luna (9/22), the Agents API public beta (9/10) and the 8/31 retirement of GPT-5.4 from Codex. On 10/01 OpenAI scheduled GPT-5.3-Codex for API removal on 2027-04-01.

A
8.3/10
Free tierFrom $2/$10
Lives inside ChatGPT -- if you already p...Parallel task execution is a real differ...
Updated 2026-11-05
Google Antigravity logo

Google Antigravity

**Gemini 4 Argon reached Antigravity on 2026-10-13 for Google AI Ultra subscribers and enterprises, and Google is offering 10,000 scientists Antigravity's highest usage tier free for six months.** Google's agent-first AI IDE -- deploys up to 5 autonomous coding agents in parallel on a VS Code fork. Antigravity 2.0 (I/O 2026) is the runtime substrate for Gemini Spark, and the Antigravity CLI is now the official successor to Gemini CLI, which stopped serving consumer tiers on 2026-06-18

A
8.0/10
Free tierFrom $0
Up to 5 autonomous agents working in par...Mission Control Manager View lets you di...
Updated 2026-10-15
Codestral (Mistral) logo

Codestral (Mistral)

Mistral's FIM-focused code-completion model -- Codestral 25.08 is a proprietary Premier API model (API id codestral-2508, GA 2025-07-30) at $0.30 in / $0.90 out per 1M tokens with a 128k context, also sold for VPC and on-prem deployment. The only open-weight Codestral is the 2024 Codestral-22B-v0.1 under the Mistral Non-Production License; older API versions codestral-2501 and codestral-2405 are retired

B
7.0/10
Free tierFrom $0.30 / $0.90
Purpose-built for IDE autocomplete -- Mi...Codestral 25.08 brought measurable upgra...
Updated 2026-11-02
Roblox Assistant logo

Roblox Assistant

**Nov 2026: Scene Generation (beta, 2026-10-30) builds editable scenes from a prompt or image, Prompt-to-Avatar entered Studio Beta on 2026-10-29, and creator-authored Skills (2026-07-09) also work in Claude Code and Cursor via Studio MCP.** Roblox Studio's agentic AI that plans, builds, and playtests games -- Planning Mode (2026-04-16), Mesh Generation and Procedural Models, free inside Studio

A
8.0/10
Free tierFrom $0
Planning Mode (2026-04-16) turns Assista...Agentic loop is genuinely real: Assistan...
Updated 2026-11-05