Now live New models for transcription and text-to-speech. AssemblyAI-STT-3.7 AssemblyAI-TTS-1
Releases & Updates

AssemblyAI-STT-3.7: for pre-recorded audio, with 32 languages and powerful speaker diarization

Replacing the Universal family of models, AssemblyAI-STT-3.7 is available now on the pre-recorded Speech-to-Text API: 32 languages (up from 18), native code switching, and a new diarization model with 36% lower DER, all at the same $0.21/hr.

AssemblyAI-STT-3.7

Written by

AssemblyAI Team

Published on

18 November 2026

Today we're releasing AssemblyAI-STT-3.7, our new flagship model for pre-recorded audio.

It nearly doubles our language coverage, from 18 to 32. It handles code switching natively, so speakers can move from Urdu to English mid-sentence and still get an accurate transcript. It also ships with a new speaker diarization model that cuts speaker mix-ups substantially. Compared with Universal-3.5 Pro at launch, cpWER is down 20%, DER is down 36% and speaker-count errors are down 63%, and it runs 3–4× faster.

The price stays the same: $0.21/hr.

Try AssemblyAI-STT-3.7 Start free, or use your existing API key.

Leading accuracy across all 32 languages, 14 of them new

Most of the world's conversations don't happen in English. If your product transcribes meetings, calls or interviews, your users expect the same quality whether they speak Hindi, Portuguese or Finnish. A model that is strong in English and uneven everywhere else forces you to choose which customers get a good product. Language coverage only counts when the accuracy comes with it.

AssemblyAI-STT-3.7 adds 14 new languages: Urdu, Russian, Korean, Catalan, Galician, Romanian, Estonian, Persian, Cantonese, Afrikaans, Marathi, Zulu, Xhosa and Norwegian Nynorsk. Several of them are already among the most accurate languages on the model: Galician at 4.1% word error rate, Russian at 4.6% and Catalan at 5.1%. See the full list of supported languages in our docs.

The languages we already supported get better too. All 17 non-English languages supported on the previous model improve, and their mean error falls from 12.8% to 10.1%.

AssemblyAI-STT-3.7's mean error falls from 12.8% to 10.1% for already supported languages

Lower is better

0 10 20 30 Spanish · Universal-3.5 Pro: 8.5% Spanish · AssemblyAI-STT-3.7: 6.9% 6.9% 8.5% Spanish Mandarin* · Universal-3.5 Pro: 9.9% Mandarin* · AssemblyAI-STT-3.7: 7.0% 7.0% 9.9% Mandarin* Danish · Universal-3.5 Pro: 10.0% Danish · AssemblyAI-STT-3.7: 7.2% 7.2% 10.0% Danish Vietnamese · Universal-3.5 Pro: 10.3% Vietnamese · AssemblyAI-STT-3.7: 7.4% 7.4% 10.3% Vietnamese Finnish · Universal-3.5 Pro: 11.4% Finnish · AssemblyAI-STT-3.7: 7.6% 7.6% 11.4% Finnish Swedish · Universal-3.5 Pro: 8.3% Swedish · AssemblyAI-STT-3.7: 7.7% 7.7% 8.3% Swedish Portuguese · Universal-3.5 Pro: 9.2% Portuguese · AssemblyAI-STT-3.7: 7.8% 7.8% 9.2% Portuguese Italian · Universal-3.5 Pro: 9.2% Italian · AssemblyAI-STT-3.7: 8.8% 8.8% 9.2% Italian Norwegian · Universal-3.5 Pro: 11.6% Norwegian · AssemblyAI-STT-3.7: 9.3% 9.3% 11.6% Norwegian German · Universal-3.5 Pro: 10.9% German · AssemblyAI-STT-3.7: 9.8% 9.8% 10.9% German Dutch · Universal-3.5 Pro: 11.2% Dutch · AssemblyAI-STT-3.7: 10.6% 10.6% 11.2% Dutch French · Universal-3.5 Pro: 12.6% French · AssemblyAI-STT-3.7: 10.7% 10.7% 12.6% French Hebrew · Universal-3.5 Pro: 13.7% Hebrew · AssemblyAI-STT-3.7: 11.1% 11.1% 13.7% Hebrew Turkish · Universal-3.5 Pro: 14.4% Turkish · AssemblyAI-STT-3.7: 11.5% 11.5% 14.4% Turkish Japanese* · Universal-3.5 Pro: 16.2% Japanese* · AssemblyAI-STT-3.7: 14.1% 14.1% 16.2% Japanese* Hindi · Universal-3.5 Pro: 27.4% Hindi · AssemblyAI-STT-3.7: 16.2% 16.2% 27.4% Hindi Arabic · Universal-3.5 Pro: 22.9% Arabic · AssemblyAI-STT-3.7: 18.1% 18.1% 22.9% Arabic
AssemblyAI-STT-3.7 Universal-3.5 Pro

Normalized word error rate; * character error rate for Japanese and Mandarin. Both models were run with the language specified in the request. Each figure is the mean over the benchmark datasets available for that language, covering read speech and conversational recordings. Languages are ordered by the AssemblyAI-STT-3.7 error rate.

Error rate on the 14 new languages

Lower is better

0 10 20 30 Galician: 4.1% Galician 4.1% Russian: 4.6% Russian 4.6% Catalan: 5.1% Catalan 5.1% Cantonese*: 6.9% Cantonese* 6.9% Korean*: 7.1% Korean* 7.1% Romanian: 9.6% Romanian 9.6% Afrikaans: 10.6% Afrikaans 10.6% Marathi: 13.0% Marathi 13.0% Urdu: 13.0% Urdu 13.0% Estonian: 13.5% Estonian 13.5% Persian: 14.1% Persian 14.1% Norwegian Nynorsk: 16.0% Norwegian Nynorsk 16.0% Zulu: 21.1% Zulu 21.1% Xhosa: 28.4% Xhosa 28.4%

Normalized word error rate; * character error rate for Korean and Cantonese. The model was run with the language specified in the request. Each figure is the mean over the benchmark datasets available for that language, covering read speech and conversational recordings.

Code switching that follows the speaker, not the accent

Supporting 32 languages is only half the story, because people don't stick to one language per conversation. And the moment they switch can be the moment that matters most.

That's exactly where many speech-to-text systems slip. Some force the whole transcript into a single language, writing Catalan as Spanish or Galician as Portuguese. Others flip a few words or a whole sentence into another language even though the speaker never switched, especially when the speaker has an accent. The real challenge is doing both at once: following every genuine switch, and staying in one language when the speaker does.

AssemblyAI-STT-3.7 does exactly that. Earlier models would sometimes transcribe accented English in another language's script; AssemblyAI-STT-3.7 keeps it in English. Across our English benchmarks, stray characters from other scripts fell by 85%, from 161 to 24 per million characters.

That did not come at the cost of real code switching. On mixed-language audio, AssemblyAI-STT-3.7 is more accurate than Universal-3.5 Pro on six of seven language pairs, with the largest gains on English–Italian, English–German and English–Arabic, where errors fell by 16% to 20%. English–French is level. The model got better at following speakers who switch, and much better at leaving alone the ones who don't.

Code switching and language flips

Lower is better

Code-switching error rate (%)

AssemblyAI-STT-3.7, mixed-language audio, by language pair

Foreign-script characters

per million characters, English audio

Error rate is normalized word error rate; * character error rate for Mandarin and Japanese. Each pair is the mean over its code-switching datasets; the change column is relative to Universal-3.5 Pro on the same datasets. Foreign-script characters are non-Latin letters in the transcripts of English benchmark audio, per million output characters. Each panel has its own scale.

Who said it matters as much as what was said

When a sentence is credited to the wrong person, everything downstream can be wrong too, even if every word is right. That's why, alongside AssemblyAI-STT-3.7, we shipped a new speaker diarization model. It makes fewer attribution mistakes, even on short utterances, gets the number of speakers right far more often, and runs 3–4× faster. It's now live for every language on all of our speech models.

Most diarization systems decide who is speaking in each stretch of audio, then give every word the label of that stretch, even where the model wasn't sure. Our new diarization model takes the model's confidence into account when assigning each word to a speaker, so sentences stay intact while quick interjections like "yeah" and "mm-hmm" are still attributed to the right speaker.

Since our previous speaker diarization model update, all three measures have improved: cpWER fell from 29.3% to 23.5%, DER from 14.7% to 9.4%, and the average speaker-count difference from 0.57 to 0.21. The gain is particularly large on short clips: on 60-second clips, cpWER fell from 34.0% to 18.6%.

Lower diarization error with AssemblyAI-STT-3.7

Lower is better

Universal-3.5 Pro + previous diarizationUniversal-3.5 Pro + new diarizationAssemblyAI-STT-3.7 + new diarization

cpWER

words plus speaker attribution

DER

diarization error rate

Speaker-count difference

predicted minus true speakers

Averages over English speaker-annotated conversation datasets. Full recordings use seven datasets; the 120-second and 60-second groups use the two of them that every system ran as clips, so the groups are not directly comparable with each other. Speaker-count difference is the predicted number of speakers minus the true number, averaged over files. Each panel has its own scale.

The improvement is not limited to English. On conversational recordings in nine other languages, cpWER is lower with AssemblyAI-STT-3.7 than with Universal-3.5 Pro in every one, falling from 23.5% to 22.1% on average.

Speaker-attributed word error rate (cpWER) by language

Lower is better

15 20 25 30 35 Italian · Universal-3.5 Pro: 19.8% Italian · AssemblyAI-STT-3.7: 18.5% 18.5% 19.8% Italian Japanese* · Universal-3.5 Pro: 21.3% Japanese* · AssemblyAI-STT-3.7: 19.9% 19.9% 21.3% Japanese* Mandarin* · Universal-3.5 Pro: 21.0% Mandarin* · AssemblyAI-STT-3.7: 20.2% 20.2% 21.0% Mandarin* Portuguese · Universal-3.5 Pro: 21.1% Portuguese · AssemblyAI-STT-3.7: 20.2% 20.2% 21.1% Portuguese Spanish · Universal-3.5 Pro: 21.3% Spanish · AssemblyAI-STT-3.7: 20.5% 20.5% 21.3% Spanish French · Universal-3.5 Pro: 24.9% French · AssemblyAI-STT-3.7: 22.0% 22.0% 24.9% French Korean (new) · AssemblyAI-STT-3.7: 22.1% 22.1% Korean (new) German · Universal-3.5 Pro: 24.4% German · AssemblyAI-STT-3.7: 23.0% 23.0% 24.4% German Turkish · Universal-3.5 Pro: 26.4% Turkish · AssemblyAI-STT-3.7: 24.4% 24.4% 26.4% Turkish Dutch · Universal-3.5 Pro: 31.5% Dutch · AssemblyAI-STT-3.7: 29.9% 29.9% 31.5% Dutch Average, 9 languages · Universal-3.5 Pro: 23.5% Average, 9 languages · AssemblyAI-STT-3.7: 22.1% 22.1% 23.5% Average, 9 languages
AssemblyAI-STT-3.7 Universal-3.5 Pro

cpWER on conversational recordings with automatic language detection on both systems. * Scored on characters. Korean is new in AssemblyAI-STT-3.7, so it has no Universal-3.5 Pro figure and is not in the average.

The diarization step also runs 3–4× faster than before. If you already set speaker_labels: true, you get it with no code change; it's live across all async speech models.

More accurate word timestamps

Word timestamps are what captions, audio and video editors, and click-to-play search are built on. A boundary that is off by a tenth of a second is enough to clip a word or show a caption late. AssemblyAI-STT-3.7 uses a new word-timing method for English, so its word boundaries line up more closely with when each word is actually spoken.

We measured the distance between each predicted word boundary and a reference timing, on regular and on noisy English recordings. At word starts, the median error fell from 57 ms to 39 ms on regular audio and from 59 ms to 40 ms on noisy audio. Word end times also improved modestly. Timestamp improvements for other languages are coming soon.

Word timestamp error on English audio

Lower is better

Word start error (ms)

0 50 100 150 200 250 Regular audio, median · Universal-3.5 Pro: 57 Regular audio, median · AssemblyAI-STT-3.7: 39 39 57 Regular audio, median Regular audio, P90 · Universal-3.5 Pro: 134 Regular audio, P90 · AssemblyAI-STT-3.7: 119 119 134 Regular audio, P90 Regular audio, P95 · Universal-3.5 Pro: 188 Regular audio, P95 · AssemblyAI-STT-3.7: 170 170 188 Regular audio, P95 Noisy audio, median · Universal-3.5 Pro: 59 Noisy audio, median · AssemblyAI-STT-3.7: 40 40 59 Noisy audio, median Noisy audio, P90 · Universal-3.5 Pro: 148 Noisy audio, P90 · AssemblyAI-STT-3.7: 130 130 148 Noisy audio, P90 Noisy audio, P95 · Universal-3.5 Pro: 225 Noisy audio, P95 · AssemblyAI-STT-3.7: 201 201 225 Noisy audio, P95
AssemblyAI-STT-3.7 Universal-3.5 Pro

Word end error (ms)

0 50 100 150 200 250 Regular audio, median · Universal-3.5 Pro: 45 Regular audio, median · AssemblyAI-STT-3.7: 40 40 45 Regular audio, median Regular audio, P90 · Universal-3.5 Pro: 127 Regular audio, P90 · AssemblyAI-STT-3.7: 120 120 127 Regular audio, P90 Regular audio, P95 · Universal-3.5 Pro: 190 Regular audio, P95 · AssemblyAI-STT-3.7: 180 180 190 Regular audio, P95 Noisy audio, median · Universal-3.5 Pro: 48 Noisy audio, median · AssemblyAI-STT-3.7: 41 41 48 Noisy audio, median Noisy audio, P90 · Universal-3.5 Pro: 149 Noisy audio, P90 · AssemblyAI-STT-3.7: 133 133 149 Noisy audio, P90 Noisy audio, P95 · Universal-3.5 Pro: 249 Noisy audio, P95 · AssemblyAI-STT-3.7: 230 230 249 Noisy audio, P95
AssemblyAI-STT-3.7 Universal-3.5 Pro

Absolute difference between predicted and reference word boundaries, in milliseconds, over all matched words: median, 90th percentile (P90) and 95th percentile (P95).

You decide how the transcript is written

Different products need different transcripts. A meeting summary needs clean prose; a legal or research record needs every "um", repetition and false start; an audio or video editing tool needs a transcript that matches the recording word for word, so that cutting a word cuts the right sound. Universal-3.5 Pro already produces high-quality clean transcripts, but since its release, customers have told us they need precise control over these aspects of transcription.

AssemblyAI-STT-3.7 exposes these controls as two independent request parameters, so you choose the style of the output:

  • disfluencies: true keeps fillers, repetitions and false starts as spoken, for verbatim transcripts (the default); false removes them for a clean transcript.
  • format_text: true writes numbers, currencies, dates and punctuation in written form (the default); false returns spoken-form text. For example:
    • Times and amounts: "twelve fifty" rather than "12:50" or "$12.50".
    • Email addresses: "jane dot doe at example dot com" rather than "jane.doe@example.com".
    • URLs: "example dot com slash pricing" rather than "example.com/pricing".
    • Phone numbers: "four one five, five five five, zero one three two" rather than "415-555-0132".

On AssemblyAI-STT-3.7 both controls default to true, giving verbatim, written-form transcripts. This is a change from Universal-3.5 Pro, where disfluencies were off by default. For clean transcripts without fillers, set disfluencies: false.

We score verbatim fidelity with four F1 metrics, each comparing the events in the transcript with those in a verbatim reference after aligning the two word by word. With disfluencies on, the model keeps most fillers and repetitions; with disfluencies off it removes them, which is what the low scores for that setting mean: the setting working as designed, not a weaker model.

Verbatim fidelity, F1 (%)

higher is more verbatim, lower is cleaner

0 20 40 60 80 100 Disfluency F1 · Disfluencies off: 34.1 Disfluency F1 · Disfluencies on (default): 75.0 34.1 75.0 Disfluency F1 Repetition F1 · Disfluencies off: 41.4 Repetition F1 · Disfluencies on (default): 72.1 41.4 72.1 Repetition F1 Contraction F1 · Disfluencies off: 33.2 Contraction F1 · Disfluencies on (default): 56.6 33.2 56.6 Contraction F1 Stutter F1 · Disfluencies off: 5.1 Stutter F1 · Disfluencies on (default): 29.9 5.1 29.9 Stutter F1
Disfluencies on (default) Disfluencies off

Mean over the benchmark datasets whose references label each event reliably. Disfluency F1 covers fillers, hesitations, backchannels and restart fragments; contraction F1 covers colloquial reduced forms such as "gonna" and "'em"; repetition F1 covers immediately repeated words; stutter F1 covers word fragments marked in the reference. Each F1 combines recall, the share of reference events the transcript keeps, and precision, the share of transcript events that are in the reference, as their harmonic mean.

Measured against the field

Beating our own previous model is only half of what you need to know. When you choose a speech-to-text provider, the question is how it compares with the alternatives. So we compared AssemblyAI-STT-3.7 with the pre-recorded models of other providers on the same recordings, each through its production API with default settings, and scored every transcript the same way, using both publicly accessible benchmarks and our own internal ones.

Public benchmarks

We start with benchmarks that anyone can inspect and reproduce: the datasets are public, and so is the scoring.

Open ASR Leaderboard

The Open ASR Leaderboard is a public benchmark with an open evaluation harness. We ran AssemblyAI-STT-3.7 through that harness and placed the result among the leaderboard's published entries.

AssemblyAI-STT-3.7 would rank 5th of 70 entries on English short-form, 3rd of 42 on multilingual and 2nd of 15 on long-form. That is a step up from Universal-3.5 Pro on every board, which ranks 16th, 4th and 6th. No single model is ahead of AssemblyAI-STT-3.7 on all three boards.

English short-form 5th of 70 mean WER over seven public sets

0 2 4 6 8 10 12 14 16 AssemblyAI-STT-3.7 · 4.93 Universal-3.5 Pro · 5.80

Ahead of it: zoom/scribe_v2.1_pro 4.19 · OtterAI/speech-v3.1 4.57 · yoob/yoob-asr-pro-v1 4.77 · microsoft/azure-speech-09-2026 4.87

Multilingual 3rd of 42 mean WER over six languages

0 2 4 6 8 10 12 14 AssemblyAI-STT-3.7 · 2.91 Universal-3.5 Pro · 3.15

Ahead of it: microsoft/azure-speech-09-2026 2.53 · elevenlabs/scribe_v2 2.76

Long-form 2nd of 15 mean WER over the long-form sets

0 2 4 6 8 10 12 14 AssemblyAI-STT-3.7 · 9.40 Universal-3.5 Pro · 10.76

Ahead of it: elevenlabs/scribe_v2 8.45

Figure 1 Where AssemblyAI-STT-3.7 would rank on the Open ASR Leaderboard (word error rate, lower is better) Each dot is one leaderboard entry; point at a dot, or tap it, for its name. For a fair comparison, every system is compared on the leaderboard's publicly available datasets only, and word error rate is computed with the leaderboard's own text normalizer. Languages whose leaderboard results mix public and private data are excluded. AssemblyAI-STT-3.7 is our own run of the leaderboard's open evaluation harness against the production API (November 2026, default settings), not an official entry; all other values are the leaderboard's published results.

AppTek call-center dialogues (Multi-Accent Long-Form Benchmark)

AppTek's call-center dialogues are a public dataset of English customer-service calls recorded across 14 accents, published with its own scoring script. We ran every system with automatic language detection and scored the transcripts with that script. With default settings, AssemblyAI-STT-3.7 has a 7.44% word error rate, close to ElevenLabs Scribe v3 at 7.34% and down from 8.35% for Universal-3.5 Pro. The dataset's references spell out every number, so we also ran the evaluation with text formatting turned off, which returns spoken-form text. With that setting, the error rate falls to 7.05%.

0 2.5 5 7.5 10 AssemblyAI-STT-3.7 (spoken-form settings): 7.05% AssemblyAI-STT-3.7 (spoken-form settings) 7.05% ElevenLabs Scribe v3: 7.34% ElevenLabs Scribe v3 7.34% AssemblyAI-STT-3.7 (defaults): 7.44% AssemblyAI-STT-3.7 (defaults) 7.44% Gemini 3.5 Transcribe: 7.52% Gemini 3.5 Transcribe 7.52% MAI-Transcribe-2: 7.71% MAI-Transcribe-2 7.71% Fish Audio Transcribe-1 Pro: 7.73% Fish Audio Transcribe-1 Pro 7.73% Grok Voice Transcribe 2.0: 7.89% Grok Voice Transcribe 2.0 7.89% Soniox stt-async-v6: 7.89% Soniox stt-async-v6 7.89% AssemblyAI Universal-3.5 Pro (previous): 8.35% AssemblyAI Universal-3.5 Pro (previous) 8.35% Deepgram Nova-3: 8.59% Deepgram Nova-3 8.59% Speechmatics Enhanced: 8.76% Speechmatics Enhanced 8.76% Voxtral Mini: 9.85% Voxtral Mini 9.85% OpenAI GPT-4o Transcribe: 10.82% OpenAI GPT-4o Transcribe 10.82%
Figure 2 Word error rate on the AppTek call-center dialogues (lower is better) 526 single-speaker recordings, about 91 hours: a 30% accent-stratified subset of AppTek's public call-center dialogues dataset · word error rate computed with AppTek's own scoring script · every system run in November 2026 with automatic language detection, using the API default settings except for the spoken-form bar · the spoken-form bar is AssemblyAI-STT-3.7 with format_text off, so numbers are spelled out as in the references, and disfluencies on · Speechmatics returned transcripts for 513 of the files and Gemini for 525; the script leaves out files that have no transcript.

Internal benchmarks

The public benchmarks above measure transcription accuracy. They do not cover speaker diarization or code switching, so for those we use our own benchmarks. AssemblyAI-STT-3.7 has the lowest error rate on both. AssemblyAI-STT-3.7 and ElevenLabs Scribe v3 were measured in November 2026; the other providers' results come from the September and October 2026 runs behind our benchmarks page.

Diarization

On speaker diarization for English conversations, AssemblyAI-STT-3.7 has the lowest diarization error rate, 10.7%, ahead of ElevenLabs Scribe v3 at 12.0%. The other systems shown are at 15% or higher.

0 10 20 AssemblyAI-STT-3.7: 10.7% AssemblyAI-STT-3.7 10.7% ElevenLabs Scribe v3: 12.0% ElevenLabs Scribe v3 12.0% Grok STT: 15.4% Grok STT 15.4% Voxtral Mini Transcribe 2: 15.9% Voxtral Mini Transcribe 2 15.9% Speechmatics Enhanced: 16.2% Speechmatics Enhanced 16.2% Gladia Solaria-1: 16.7% Gladia Solaria-1 16.7% AWS Transcribe: 16.9% AWS Transcribe 16.9% Deepgram Nova-3: 17.4% Deepgram Nova-3 17.4% Soniox stt-async-v5: 24.2% Soniox stt-async-v5 24.2%
Figure 3 Average diarization error rate (lower is better) Diarization error rate (DER), averaged over six English conversation datasets · AssemblyAI-STT-3.7 and ElevenLabs Scribe v3 measured November 2026; all other systems October 2026 · each vendor's production pre-recorded API with default settings.

Code switching

Code switching is where the field spreads out most. AssemblyAI-STT-3.7 has the lowest average error rate, 6.2%, on audio that mixes English with German, Spanish, French, Italian or Portuguese, and the lowest error rate on 12 of the 16 datasets. ElevenLabs Scribe v3 follows at 6.8% and leads on the other four; Soniox is at 10.2%. The remaining systems are at 17% or higher, and several are above 40%.

0 20 40 AssemblyAI-STT-3.7: 6.2% AssemblyAI-STT-3.7 6.2% ElevenLabs Scribe v3: 6.8% ElevenLabs Scribe v3 6.8% Soniox stt-async-v5: 10.2% Soniox stt-async-v5 10.2% Azure Batch Transcription: 17.4% Azure Batch Transcription 17.4% Voxtral Mini Transcribe 2: 17.4% Voxtral Mini Transcribe 2 17.4% AWS Transcribe: 22.9% AWS Transcribe 22.9% Parakeet TDT 0.6B v3: 29.3% Parakeet TDT 0.6B v3 29.3% Cohere Transcribe (language pinned): 29.4% Cohere Transcribe (language pinned) 29.4% OpenAI GPT-4o Transcribe: 44.7% OpenAI GPT-4o Transcribe 44.7% Deepgram Nova-3: 49.4% Deepgram Nova-3 49.4% Gladia Solaria-1: 50.4% Gladia Solaria-1 50.4%
Figure 4 Average word error rate for code-switching speech recognition (lower is better) English paired with German, Spanish, French, Italian and Portuguese · normalized word error rate · AssemblyAI-STT-3.7 and ElevenLabs Scribe v3 measured November 2026; all other systems September 2026 · each vendor's production pre-recorded API with default settings.

Pricing

AssemblyAI-STT-3.7 is available for $0.21/hr of audio, unchanged from the previous model. Your current rate for pre-recorded API access stays the same. See our pricing.

Get started

AssemblyAI-STT-3.7 is available now with your existing API key. Change the model name, run your own recordings through it, and start seeing improvements today.

Try it on your own audio Use your existing API key.

Frequently asked questions

How do I switch from Universal-3.5 Pro?

Set speech_models to ["assemblyai-stt-3-7"]. There is no new endpoint. For the full steps, see the AssemblyAI-STT-3.7 migration guide.

import requests

response = requests.post(
    "https://api.assemblyai.com/v2/transcript",
    headers={"authorization": "YOUR_API_KEY"},
    json={
        "audio_url": "https://example.com/meeting.mp3",
        "speech_models": ["assemblyai-stt-3-7"],
        "speaker_labels": True,
    },
)
print(response.json()["id"])

If I don't set speech_models, will I get the latest model?

It depends on your account:

  • New accounts created on or after November 17, 2026: requests that don't set speech_models already use assemblyai-stt-3-7.
  • Existing accounts that default to universal-3-5-pro or universal-3-pro: starting December 8, 2026, requests that don't set speech_models will use assemblyai-stt-3-7 automatically. This changes the default model only; universal-3-5-pro and universal-3-pro are not being retired on that date.
  • Existing accounts that default to universal-2: these are not upgraded automatically.

If you do set speech_models, your requests keep using the model you specified until that model is retired. To test AssemblyAI-STT-3.7 and switch to it, see the AssemblyAI-STT-3.7 migration guide.

What about languages outside the 32?

Set speech_models: ["assemblyai-stt-3-7", "universal-2"] for support across 99 languages. If you set language_code to a language AssemblyAI-STT-3.7 doesn't support, the request goes to Universal-2. The speech_model_used field tells you which model ran. See transcribe all 99 languages by combining speech models in our docs.

Is Universal-3.5 Pro still available?

Universal-3 Pro and Universal-3.5 Pro are deprecated, with a retirement date to be announced. Once they are retired, requests specifying universal-3-pro or universal-3-5-pro will route to AssemblyAI-STT-3.7 automatically. Universal-2 remains available. See model versions and deprecations in our docs.