Is Microsoft MAI-Transcribe-1.5 Down?

Quickest ways to check the current status of microsoft.ai plus recent known issues and working alternatives if it's out.

Last editorial review: 2026-10-02

How to check right now

Step 1 -- Load the site
Open microsoft.ai→

The simplest check. If it loads and you can log in, the tool is up for you.

Step 2 -- Crowdsourced reports
DownDetector: Microsoft MAI-Transcribe-1.5→

User-reported outages aggregated across regions. Good signal for partial or regional outages.

Step 3 -- Live social
Search X for “Microsoft MAI-Transcribe-1.5down” →

Live tweets from other users. Fastest signal for a fresh outage -- often beats status pages by 10-20 minutes.

Step 4 -- Official status page
Find Microsoft MAI-Transcribe-1.5's status page →

Most vendors publish a status page at status.[domain] or on Statuspage.io. Google-search to find the current URL.

Known issues we've tracked

MAI-TRANSCRIBE-2-STREAMING -- MICROSOFT'S FIRST STREAMING ASR, NO. 1 ON ARTIFICIAL ANALYSIS FOR PARTIALS AND FINALS, $0.54 PER HOUR INTRO (2026-10-01, vendor-primary): Microsoft AI launched **MAI-Transcribe-2-Streaming**, which 'delivers low-latency, real-time transcripts in 60 languages, all while supporting automatic, continuous language detection'. **Leaderboard claim, as a dated snapshot:** 'It ranks no. 1 for accuracy for both final and partial transcripts on Artificial Analysis', and on the accuracy-versus-latency evaluation Microsoft says it 'sit[s] on the Pareto frontier'. **Mechanics:** rather than waiting for the speaker to finish, it 'produces its first hypotheses (known as partials) in just over 100ms of receiving audio', revises them, and commits a stable transcript; Microsoft's internal evaluation says words appear 'in the transcript 2x faster than with our closest competitor' for dictation and subtitling, and voice agents 'can start reasoning or calling tools mid-sentence'. **Price: 'available at an introductory price of $0.54 per hour of audio through the end of the year'** -- a dated flip (post-2026 rate unstated) and 5.4x the batch MAI-Transcribe-2 rate of $0.10/hr recorded below. **Launched alongside MAI-Voice-2.1 ($22 per 1M characters, 23 languages / 26 locales, one voice across languages with a native accent) and MAI-Voice-2.1-Flash ($15 per 1M characters, 45 s of audio at 150 ms end-to-end latency)**, which Microsoft positions as the other half of a voice-agent loop; a 'Chatter' demo in the MAI Playground shows them together. The voice models are on OpenRouter; the post's distribution list for all three was cut off in our extract, so Azure Foundry availability is assumed from the Transcribe-2 precedent rather than verified. Comparison for this page: Grok Voice Transcribe 2.0 is $0.10/$0.20 per hour (streaming), Gemini 3.5 Transcribe about $0.005/min ($0.30/hr) -- so $0.54/hr is a premium streaming price justified by the accuracy claim, not a price lead.
2026-10-01•Microsoft AI (microsoft.ai/news/our-first-streaming-transcription-model/, JSON-LD datePublished 2026-10-01T16:00Z; listed first on microsoft.ai/blog/ via Firecrawl) -- fetched 2026-10-02 via curl with browser UA
MAI-TRANSCRIBE-2 SUPERSEDES THIS MODEL -- AND THE PRICE IS THE HEADLINE, NOT THE ACCURACY (2026-09-03, vendor-primary): Microsoft launched **MAI-Transcribe-2**, calling it 'not only our most capable transcription model yet, but the most capable and efficient amongst our competitors'. **THE NUMBER THAT MATTERS: $0.10 per hour of audio**, which Microsoft calls 'the best price on the market'. **NEW CAPABILITIES OVER 1.5:** **diarization** (speaker separation), **configurable transcription styles**, and **word-level timestamps** -- the diarization addition is the significant one, because its absence is what pushed transcription buyers toward competitors regardless of raw WER. **ACCURACY, VENDOR-PUBLISHED:** ranks **first on the FLEURS benchmark across 60 languages with an average Word-Error-Rate of 5.2%**, defines the Pareto frontier for accuracy and latency on **Artificial Analysis**, and ranks **second on the Artificial Analysis WER leaderboard**. Microsoft names the models it says it beats: **Gemini 3.5 Transcribe, GPT-Transcribe, Whisper V3-Large and ScribeV2**. **READ THE TWO WER FIGURES CAREFULLY BEFORE COMPARING THEM TO THIS PAGE'S 1.5 NUMBERS -- THEY ARE DIFFERENT BENCHMARKS.** The 5.2% is a **60-language FLEURS average**; the 2.4% recorded on this page for MAI-Transcribe-1.5 is an **Artificial Analysis** figure. A higher headline number across sixty languages is not a regression against a narrower English-weighted benchmark, and Microsoft claims a rank improvement on Artificial Analysis (No. 3 for 1.5, No. 2 for 2) rather than a decline. **Language coverage also expands: 43 languages on 1.5, 60 evaluated on 2.** **STATUS OF THIS PAGE'S MODEL:** Microsoft's own model index now lists MAI-Transcribe-2 as the current transcription model; **no retirement date has been published for MAI-Transcribe-1.5**, so this page is not marking it deprecated -- but new deployments should be evaluated against 2, and the 1.5 integrations into Copilot, Teams, GitHub and Dynamics 365 Contact Center should be expected to migrate.
2026-09-03•Microsoft AI (microsoft.ai/news/mai-transcribe-2-is-the-fastest-most-accurate-and-cheapest-speech-recognition-model-in-the-world/, JSON-LD datePublished 2026-09-03) -- fetched 2026-09-05 via curl with browser UA
VERSION BUMP (2026-06-02, Microsoft Build): MAI-Transcribe-1.5 launched in the 'seven new MAI models' wave. Vendor-published changes vs 1.0: language coverage expanded from 25 to 43; speed improved to under 15 seconds per hour of audio (from 53s); new keyword-biasing feature delivers a ~30% WER reduction on the FLEURS multilingual benchmark by applying domain terminology intelligently (context-aware, not blind keyword forcing). Microsoft claims 'best-in-class Word Error Rate across 43 languages' and #3 on the Artificial Analysis leaderboard at 2.4% WER, describing it as 'the fastest, most efficient and most cost-effective transcription model of any hyperscaler.' Now being integrated into Copilot, Teams, GitHub, and Dynamics 365 Contact Center; available through Foundry. Per-hour pricing not disclosed at launch.
2026-06-02•Microsoft AI (microsoft.ai/news/mai-transcribe-1-5more-accurate-context-aware-and-built-for-production/, microsoft.ai/news/building-a-hillclimbing-machine-launching-seven-new-mai-models/)
Public preview in US only for MAI Playground. Foundry API works globally but you need an Azure subscription to evaluate
2026-04•Microsoft AI launch post
Competitor positioning on the site: MAI-Transcribe-1 is a backend model, not a meeting-transcription product. Do not position it as an Otter.ai competitor -- it competes with Whisper and would typically be adopted BY meeting apps, not replace them
2026-04•Microsoft model card + tech analysis

Issues here are sourced from our editorial sweeps, not real-time telemetry. Newer issues may exist.

What to use if Microsoft MAI-Transcribe-1.5 is down

Top AIToolTier-ranked alternatives in the same category, ordered by our overall score.

B

Fireflies.ai

7.8

AI meeting notetaker with 6,000+ integrations that records, transcribes, and summarizes your calls **Pricing refreshed 2026-10-06: Free now stores 400 minutes per team (was 800) with 20 AI credits; Pro is $18 monthly or $10 per seat on annual billing; Business $29 or $19 with unlimited storage; a new Enterprise tier at $39 per seat, annual only; plus Fireflies Talk, a free unlimited dictation tool.**

B

Otter.ai

7.5

Joins your meetings, transcribes everything, and gives you a summary so you can actually pay attention. **Pricing re-verified 2026-10-05: Basic stays free at 300 minutes/month with a 30-minute cap per conversation; Pro is $16.99/user/month monthly and annual billing drops Pro to $8.33; Business is $30 monthly or $19.99 annual and now advertises unlimited meetings and imports instead of a 6,000-minute cap.**

B

Notion AI

7.0

**Notion's 2026-10-13 release lets Business and Enterprise members connect a ChatGPT Plus or Pro subscription to cover their Notion AI usage with no markup, adds a Skills API that syncs Skills to ChatGPT and Claude, column-level database permissions, admin controls for MCP-connected agents, rebuilt GitHub/Jira/Salesforce synced databases and seven new models including GPT-6.1 Sol and Opus 5.5.** **Notion 3.7 (2026-09-15)**: Agent skills for the whole team -- reusable instructions that teach Notion Agent how your team works, kept in a shared skills library and run with a slash command. **Model controls (2026-09-09)**: Business/Enterprise workspace owners can now pick which AI models Notion Agent and Custom Agents may use and set a default -- a cost and compliance lever the page lacked. Also new since July: suggest-edits mode (8/28), a model picker with scorecards (8/14), the Developer Portal in the sidebar (8/19), and AI Meeting Notes triggering Custom Agents (7/31). AI built into Notion -- Custom Agents exited beta 2026-05-04/05 with admin controls (per-agent spend caps, workspace caps, auto-pause on limit) and credit billing live at $10 / 1,000 credits (~$0.17-$0.33 per agent run) as an add-on to Business/Enterprise

About Microsoft MAI-Transcribe-1.5

Tier B (7.9/10). Microsoft's in-house speech-recognition line. **MAI-Transcribe-2-Streaming launched 2026-10-01 -- Microsoft's first real-time streaming transcription model, ranked No. 1 on Artificial Analysis for both final and partial transcripts, 60 languages with continuous language detection, first partial hypotheses just over 100 ms after audio arrives, at an introductory $0.54 per hour of audio through the end of 2026.** It sits beside the batch **MAI-Transcribe-2 (2026-09-03, $0.10 per hour)**, which superseded the 1.5 model this page was built on: diarization, configurable styles, word-level timestamps, No. 1 on FLEURS across 60 languages. Microsoft pairs the streaming model with the new MAI-Voice-2.1-Flash for voice agents.

Read the full Microsoft MAI-Transcribe-1.5review →