ChatGPT Pricing
All plans and pricing as of 2026-10-02
Pro 500 (new 2026-09-29)
- ✓Highest included usage of the Pro plans -- 25 times the ChatGPT Plus allowance per the DevDay recap
- ✓**Access to Astra Ultrafast in ChatGPT Work and Codex** -- a faster service tier for GPT-6 Astra (up to 8x faster token generation in Codex); among Pro plans only Pro 500 unlocks it, and buying credits on Pro 100 or Pro 200 does not
- ✓Ultrafast uses the included allowance first, then the credit balance
- ✓Web-only subscriptions at launch
- ✓Agents API with computer use in Codex and ChatGPT Work
API -- GPT-6.1 Sol (launched 2026-09-29)
- ✓Cache writes $2.50; above 272K input the request bills $4 input / $0.20 cached / $5 writes / $15 output
- ✓Fast mode 2x Standard; Batch and Flex 50% lower; regional processing +10%
- ✓1,050,000 context, 922,000 max input, 128,000 max output, Apr 30 2026 knowledge cutoff, text + image in
- ✓reasoning.effort low / medium (default) / high / xhigh / max -- none and minimal are not supported
- ✓Responses API for tool calling; Chat Completions without tools; no fine-tuning, no Live/Realtime
- ✓GPT-6.1 Sol Ultrafast 'in the coming days'; GPT-6 Astra Ultrafast on the API is $60/$300 per 1M
Free
- ✓GPT-6 Luna in ChatGPT Work and Codex via the desktop app from 2026-09-22 (Chat itself stays on GPT-5.6 Luna -- GPT-6 models are 'not yet available in Chat')
- ✓**GPT-5.6 Luna becomes the default model this week (announced 2026-08-06)** -- a jump of several model generations for free users
- ✓**UNLIMITED text chats starting the week of 2026-08-10** (subject to abuse guardrails); limits still apply to file uploads, images and other tools
- ✓New **Think button** gives Luna more time on harder questions
- ✓Web browsing
- ✓Basic image gen
Go
- ✓**GPT-5.6 Luna default + unlimited text chats + Think button on the same 2026-08-06 announcement as Free**
- ✓Codex with GPT-5.5 at 400K context
- ✓Image generation
- ✓File uploads
Plus
- ✓GPT-6 Sol + GPT-6 Luna in ChatGPT Work and Codex (2026-09-22; not in Chat)
- ✓GPT-5.5 Thinking (new 2026-04-23)
- ✓GPT-5.4 remains available
- ✓Custom GPTs
- ✓1M token context
- ✓Image gen via ChatGPT Images 2.5 (2026-09-08; Sketch, templates, up to 50% faster than Images 2.0)
Pro ($100, new 2026-04-09)
- ✓5x Plus usage (10x Codex during launch promo through 2026-05-31)
- ✓GPT-5.5 Thinking + GPT-5.5 Pro (Pro/Business/Enterprise only)
- ✓Targets Claude Max price point
- ✓Best for longer high-effort Codex sessions
Pro (Ultra $200)
- ✓GPT-5.5 Pro (launched 2026-04-23)
- ✓Unlimited usage
- ✓Priority access
- ✓Power users and agent workloads
Business -- Standard seat
- ✓GPT-5.5 Thinking + GPT-5.5 Pro
- ✓Admin controls
- ✓Data privacy
- ✓Custom GPTs
- ✓Subject to the five-hour usage limit that Premium removes
Business -- Premium seat (ANNOUNCED 2026-08-10, STILL WAITLIST as of 2026-08-20 -- not yet GA)
- ✓**5x more usage than Standard**
- ✓**No five-hour usage limit**
- ✓Predictable weekly usage resets
- ✓Mix Standard and Premium seats in the same workspace; upgrade/reassign as needs change
- ✓Shared workspace credits available if you still hit a limit
- ✓**STILL NOT generally available -- re-verified on the vendor page 2026-08-20, ten days after announcement: the post still reads 'Premium seats are coming', still offers only 'Join the waitlist', and still gives no GA date**
- ✓Launch promo: $100 in workspace credits (2,500) per Premium seat, max $500 / 5 seats, first 10,000 eligible customers -- **the promo CLOSED 2026-08-20 and the seats never went GA inside the promo window**, so anyone who joined the waitlist for the credits is holding an unredeemable offer until OpenAI ships
API -- GPT-6 Sol / GPT-6 Luna (launched 2026-09-22)
- ✓Cached input $0.20 (Sol) / $0.01 (Luna); cache writes $2.50 / $0.125; prompts above 272K input billed 2x input and 1.5x output
- ✓1,050,000 context, 922,000 max input, 128,000 max output; cutoffs Apr 20 (Sol) / May 18 (Luna), 2026
- ✓Batch and Flex at 50%, Fast mode at 2x; EU data residency Standard-only; Bedrock at OpenAI list prices
- ✓Improved prompt caching: 30-minute reuse window, 90% cached-read discount, dashboard + diagnostics, effort changes no longer break cache
API -- GPT-5.6 family (GA 2026-07-09; Terra + Luna CUT 2026-07-30)
- ✓PRICE CUT 2026-07-30: Luna -80% ($1/$6 -> $0.20/$1.20), Terra -20% ($2.50/$15 -> $2/$12); Sol unchanged
- ✓Fast mode replaces Priority Processing: Sol up to 2.5x faster at 2x price, same intelligence
- ✓GA across ChatGPT/Codex/API 2026-07-09
- ✓Programmatic Tool Calling (ZDR-compatible) + multi-agent beta in Responses API
- ✓Prompt caching: 90% read discount, 30-min minimum cache life, cache WRITES now 1.25x uncached input
- ✓max/ultra effort settings (ultra = 4 parallel agents)
API -- gpt-5.5
- ✓LIVE as of 2026-04-24 (shipped ~24h after consumer launch)
- ✓Cached input $0.50/M (90% discount)
- ✓Batch + Flex = 50% of standard
- ✓Priority = 2.5x standard
- ✓1M context window
API -- gpt-5.5-pro
- ✓LIVE as of 2026-04-24
- ✓1M context window
- ✓Most expensive frontier LLM API on the market today
Is ChatGPT Worth the Price?
Value Score: 8/10
Overall Score: 8.8/10 · Everyone. Seriously -- if you're new to AI or want the most complete all-in-one package, ChatGPT is the default recommendation.
ChatGPT is the Toyota Camry of AI chatbots. It's not the flashiest or the absolute best at any one thing, but it does everything well, it's reliable, and the ecosystem around it is massive. The July 9, 2026 wave is the biggest single week in the product's history: GPT-5.6 GA (Sol/Terra/Luna, with `ultra` multi-agent mode), GPT-Live full-duplex voice, the ChatGPT Work agent, and a unified desktop app free on every plan. OpenAI's pitch shifted from raw capability to performance-per-dollar -- and on July 30 it made that literal, cutting Luna 80% to $0.20/$1.20 and Terra 20% to $2/$12 while leaving Sol alone. On August 6 it pushed that logic down to the free tier: Luna by default, unlimited text chats, and a Think button, which makes the free plan genuinely hard to beat. Paid users got ergonomics rather than capability -- one model and an effort slider instead of the Instant/Thinking split -- and that retune stops at the Chat surface, so Codex and Work behave exactly as they did. For most people, this is still the only AI chatbot they need, and the gap to needing anything else just narrowed again.
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.
How ChatGPT Pricing Compares
| Tool | Free Tier | Starting Price | Value Score | Overall |
|---|---|---|---|---|
| ChatGPT(this tool) | Yes | $500/month | 8/10 | 8.8 |
| Perplexity AI | Yes | $0 | 9/10 | 8.5 |
| Alexa+ | Yes | $0 | 8.5/10 | 8.3 |
| Gemini Notebook (formerly NotebookLM) | Yes | $0 | 9.5/10 | 7.8 |
| Microsoft Copilot | Yes | $0 | 8/10 | 7.5 |
| Poe | Yes | $0 | 8/10 | 7.5 |