# Is AssemblyAI-STT-3.7 SOTA?

Pending benchmarks. Announced Nov 17, 2026; no independent results yet.

As of Nov 22, 2026, AssemblyAI-STT-3.7 is pending benchmarks on isitsota: announced Nov 17, 2026, with no independent results on the boards we track yet.

AssemblyAI's pre-recorded speech-to-text model in the AssemblyAI-STT-3.7 family, for transcripts of recorded audio and video. Official docs describe 32 languages with automatic language detection and native code switching, plus a new speaker diarization model. It replaces Universal-3.5 Pro for pre-recorded audio. Live on AssemblyAI's pre-recorded Speech-to-Text API as speech_models: ["assemblyai-stt-3-7"] at $0.21 per hour of audio. No weights. Official models.md dates the family launch to November 17, 2026; the changelog and lab blogs are dated November 18.

Canonical: https://isitsota.com/m/assemblyai-stt-3-7
Last updated: 2026-11-18
Tracking since: 2026-11-18

## Announced facts

| Fact | Value |
| --- | --- |
| Lab | AssemblyAI |
| Announced | Nov 17, 2026 |
| API | Live (`assemblyai-stt-3-7`) · Pre-recorded Speech-to-Text API: POST https://api.assemblyai.com/v2/transcript with speech_models: ["assemblyai-stt-3-7"] and an API key. There is no new endpoint. Official docs: new accounts created on or after November 17, 2026 already default to this model when speech_models is omitted; existing accounts that default to universal-3-5-pro or universal-3-pro switch on December 8, 2026 unless they set the model explicitly. Universal-2 defaults are not switched automatically |
| Open weights | Closed · Not released |
| Task | Pre-recorded speech-to-text for meetings, calls, podcasts, and other recorded audio or video |
| Languages | 32 languages with automatic language detection and native code switching. Official research post says coverage rose from 18 to 32, adding Urdu, Russian, Korean, Catalan, Galician, Romanian, Estonian, Persian, Cantonese, Afrikaans, Marathi, Zulu, Xhosa, and Norwegian Nynorsk. For the remaining languages of the 99 Universal-2 covers, set speech_models to ["assemblyai-stt-3-7", "universal-2"] |
| Pricing | $0.21 per hour of audio, the same rate AssemblyAI states for the previous flagship pre-recorded model; volume discounts are available. Official pricing also lists an optional Medical Mode add-on at +$0.15/hr |
| Availability | Generally available on AssemblyAI's self-serve pre-recorded Speech-to-Text API. Official changelog dated 2026-11-18. Universal-3 Pro and Universal-3.5 Pro are deprecated as of November 17, 2026, with a retirement date to be announced; after retirement, requests naming them route to this model |
| Related | Separate from AssemblyAI-STT-3.7-Realtime and AssemblyAI-STT-3.7-Realtime-Agent, which are streaming models, from AssemblyAI-TTS-1 (text-to-speech), and from Universal-3.5 Pro |

Tags marked unsure are not stated on the lab's own pages.

## What would let us call it

Verdicts use only independent boards we already track. Two consecutive snapshots are required before any crown.

- [x] Publicly announced with a source
- [x] Usable by the public (API or weights)
- [ ] Listed on an independent board we track
- [ ] Enough results for a rank in a category
- [ ] Holds across two snapshots (crowns only)

## What the lab claims

Self-reported. Shown for context; not used in any verdict.

- Lab claim: compared with Universal-3.5 Pro at launch, cpWER is down 20%, diarization error rate (DER) is down 36%, and speaker-count errors are down 63%, and it runs 3–4× faster (self-reported, official AssemblyAI-STT-3.7 research post).
- Lab claim: on the 17 non-English languages already supported, mean error fell from 12.8% to 10.1%; among the 14 new languages, AssemblyAI reports Galician at 4.1% word error rate, Russian at 4.6%, and Catalan at 5.1% (self-reported, official research post).
- Lab claim: across AssemblyAI's English benchmarks, stray characters from other scripts fell by 85%, from 161 to 24 per million characters (self-reported, official research post).

## Sources

- LAB: [AssemblyAI models reference (llms/models.md)](https://www.assemblyai.com/llms/models.md) (2026-11-17)
- LAB: [Speech model versions and deprecations](https://www.assemblyai.com/docs/getting-started/model-versions-and-deprecations) (2026-11-17)
- LAB: [AssemblyAI changelog: AssemblyAI-STT-3.7](https://www.assemblyai.com/changelog/) (2026-11-18)
- LAB: [Introducing AssemblyAI-STT-3.7 (AssemblyAI blog)](https://www.assemblyai.com/blog/introducing-assemblyai-stt-3-7) (2026-11-18)
- LAB: [AssemblyAI-STT-3.7: for pre-recorded audio (research post)](https://www.assemblyai.com/blog/assemblyai-stt-3-7-research) (2026-11-18)
- LAB: [assemblyai-stt-3-7 migration guide](https://www.assemblyai.com/docs/pre-recorded-audio/assemblyai-stt-3-7-migration-guide)
- LAB: [AssemblyAI pricing reference](https://www.assemblyai.com/pricing.md)
- LAB: [AssemblyAI on X: AssemblyAI-STT-3.7](https://x.com/AssemblyAI/status/2123080177508929849) (2026-11-18)

Tracking continues until an independent board lists it.

## Related

- Lab: [AssemblyAI](https://isitsota.com/labs/assemblyai.md)
- All pending releases: https://isitsota.com/releases.md
- JSON: https://api.isitsota.com/v1/models/assemblyai-stt-3-7
