MAI-Transcribe-2 at $0.54 an hour vs Sume STT at 1 cent a minute
MAI-Transcribe-2-Streaming is $0.54 an hour as an intro price through 2026. Sume STT is $0.01 a minute, $0.60 an hour. The bill at 10, 100 and 1,000 hours.

Microsoft's MAI-Transcribe-2-Streaming launched on 2026-10-01 at an introductory $0.54 per hour of audio through the end of 2026, with 60 languages (as reported by Digital Applied and PYMNTS, read 2026-10-06). Sume's speech-to-text is billed per audio minute at $0.01 catalog price, which is $0.60 an hour. Neither source gives the post-2026 price of MAI-Transcribe, so this post compares only the intro rate.
Monthly bill by volume
| Hours a month | MAI-Transcribe-2 (intro) | Sume STT, catalog | Sume STT, with 5.5% fee | Sume catalog minus MAI |
|---|---|---|---|---|
| 10 | $5.40 | $6.00 | $6.33 | $0.60 |
| 100 | $54.00 | $60.00 | $63.30 | $6.00 |
| 1,000 | $540.00 | $600.00 | $633.00 | $60.00 |
What differs besides price
MAI-Transcribe-2-Streaming is a streaming model: you get text while the audio is still playing. Sume STT is an asynchronous job on an uploaded recording, so it fits transcripts for editing, captions and search rather than live captions. If you need live captions, the Microsoft model is built for that and Sume's is not.
The difference at 100 hours a month is about $6. For recorded media, the more likely deciding factor is whether you want the transcript to feed straight into Sume captions, trims and dubs, where the same job ids and balance apply. See 40 meeting hours a month for another worked case.
Check the date
Because the Microsoft price is an intro rate, rerun this table in January 2027 with Microsoft's own price page. Sume's STT price is a catalog constant; both numbers appear in the catalog your key can read.
Sources
Related posts
More in Comparisons
- MAI-Voice-2.1 or Sume TTS? Three questions that decide it
Live speech, extra outputs or lowest price per character? A short guide to MAI-Voice-2.1 and Flash against the Sume TTS Router, with 10M-character math.
- MiniMax H3 open weights vs a hosted lip-sync API: what you take on
MiniMax released H3 open weights on 2026-08-03. Self-hosting is not the same as a hosted still-plus-audio lip-sync route. What each choice makes you own.
- Nova Canvas 4,194,304 pixel cap and 16 px rule vs Sume image_size
Nova Canvas output sides must divide by 16 and total under 4,194,304 pixels. GPT Image 2.5 on Sume has its own custom-size rules. A side-by-side check.
- Nova Canvas IMAGE_VARIATION similarityStrength vs Sume n and reference
Nova Canvas IMAGE_VARIATION takes 1 to 5 images and a similarityStrength of 0.2 to 1.0. Sume has no strength field; here is how to get variations.
Written by Sume