MAI-Transcribe-2
Among the most accurate transcription models at $0.10 an hour. Still a preview with no SLA, so not for production yet.
Should you use it?
Pick it if
Developers transcribing lots of recorded audio who can live with preview status.
Best for: Developers transcribing lots of recorded audio who can accept a preview (Speech to text Top 20)
Skip it if
You need a supported production service today; ElevenLabs Scribe v2 is close in accuracy and not labelled a preview.
You give up: Preview with no service guarantee; price after 2026 not announced
The short version
- What this is
- Microsoft's speech to text model. It turns recorded audio into text in 60 languages through Microsoft Foundry, at a low introductory price.
- Why it matters
- It is one of the most accurate transcription models available and among the cheapest per hour.
- 2.0% word error rate, 3rd of 55 models on Artificial Analysis; no score yet, as its terms stop us publishing one.
- $0.10 per audio hour until the end of 2026.
- Public preview: no SLA.
Good at
Our take
On Artificial Analysis's word error rate test, only two models make fewer mistakes than MAI-Transcribe-2, and both are slower s1. At $0.10 an hour it is also among the cheapest strong options s2.
The catch is status. Microsoft lists it as a public preview without an SLA and advises against production use s4, and the price is introductory until the end of 2026 s3. Microsoft says it ranks second on Artificial Analysis s3; when we sorted the current table it was third s1.
Score
A score from 0 to 100 per job, against everything available today.
New to this? What the scores and labels mean
- 0 to 100
- Our score for one job, like writing or coding, compared with everything else out today. 100 is the best available today. For AI models we build it from public data, such as Arena's blind votes and Epoch AI's independent tests; apps and tools are scored on our own checklist. The same tool can score differently for different jobs, and where two likely ranges overlap, we call it a tie.
- Our verdict
- Use it: a top choice for most people. Worth trying: good, and worth a go. Depends: good for some people or one kind of job. Wait: too early to tell, or not out for everyone yet. Skip: something else does the job better.
- First look
- Just released. We say what it is and whether to bother, and give it a score once public results exist.
- No score yet
- None of the public sources we use covers it yet, or the only evidence is from a source whose numbers we can't use, so we describe it in words instead.
- Best buy
- Nothing cheaper scores as well for that job.
- Best free
- The best score for that job that costs nothing, through a free plan or by running on your own computer.
- Pay as you go
- No monthly plan: you pay for each use, usually through a developer account.
- Checked
- The day we last confirmed the score and the price against the sources. Prices change often, so check the date.
- AI model
- The engine underneath, like GPT or Claude. You use it through an app or, if you code, directly.
- Open source
- Free to download and run on your own computer, so nothing leaves your machine.
Voice and audio
No score yet. The leaderboard this rests on does not let us publish numbers from it, so we describe the evidence instead.
1 source so far
Independent testArtificial Analysis speech to text: word error rate 2.0%, rank 3 of 55 distinct models (59 rows, some the same model from different hosts) when sorted by error rate, lower is better; two models tie ahead at 1.7% s1.
Alternatives
Nothing we have scored beats it for voice on score or on price, and we have not scored a free option for it yet.
- Scribe v2Voice model–FreeClose in accuracy, not labelled a preview, with speaker labels and a free plan. Side by side
- Gemini 3.5 TranscribeVoice modelNot reviewed yetError rate 2.6% on Artificial Analysis, from Google's API.
- Grok Voice Transcribe 2.0Voice modelNot reviewed yetError rate 2.3% on Artificial Analysis at the same $1.67 per 1,000 minutes s1.
Evidence
Everything behind the call and the scores, folded so the page stays short. Spotted a wrong price, score or link? Report a mistake.
How we scored this
Voice has no score yet, because none of the public sources we use covers it.
Independent test
Reviews are drafted with help from AI models, including Anthropic's Claude. Model scores come only from outside leaderboards and independent evaluations, by a written rule.
No one paid for this review or gave us free access, and we earn nothing if you sign up. How we score
Sources 5
- s1Artificial Analysis speech to text leaderboard
- s2Microsoft AI: MAI-Transcribe-2
- s3Microsoft AI news: MAI-Transcribe-2 announcement
- s4Microsoft Learn: MAI-Transcribe in Azure Speech (preview)
- s5Microsoft Learn: speech to text data, privacy and security
Something here out of date or wrong? Report a mistake and we will check it against the source.
History 2 changes
Every change to a rating, the price or our call, with the reason. Scores are against what exists on the day, so they drift down as better things ship. Older entries come from before we moved to scores.
- New: Voice ratedReviewed: voice.
- Audit: re-opened Artificial Analysis speech to text; 59 rows, 2.0% rank 3 behind two 1.7% models confirmed. Replaced gpt-transcribe (3.3%, pricier) with Grok Voice Transcribe 2.0 (2.3%, same price) as an alternative. Rating unchanged.
Use it well 4 guides on this site
Pages on this site that help with the jobs above.