STT.ai Enhanced
STT.ai runs this model on the paid tier. It is the model behind STT.ai Enhanced. Free transcription uses Whisper Large v3 Turbo, which is faster.
See plans →
·
Transcribe free with Whisper Turbo →
4.2%
WER
99
Languages
15.9x
Speed
MIT
License
About STT.ai Enhanced
STT.ai Enhanced is our name for Whisper large-v3 running on STT.ai's GPU. It is not a custom model or a fine-tune — it is OpenAI's full 1.55B-parameter Whisper checkpoint, the most accurate of the Whisper family, with 99 languages and speaker diarization. It is more accurate than the free default (Whisper large-v3-turbo, an 809M distillation) and roughly 1.5x slower on a diarized job: measured on our hardware, 300 seconds of speech takes 18.8s with Enhanced versus 13.0s with Turbo.
✦ Unlock Enhanced Model
Get access to our most accurate model with any paid plan — Whisper large-v3, 99 languages, and speaker diarization.
View Plans →Model Info
- ProviderOpenAI (hosted by STT.ai)
- Architecture-
- LicenseMIT
- UpdatedAug 2026
Frequently Asked Questions
STT.ai Enhanced is a speech-to-text model by OpenAI (hosted by STT.ai). STT.ai runs it on our own GPU for accounts on a paid plan, and it is the model behind the STT.ai Enhanced tier. Free transcription uses Whisper Large v3 Turbo, the faster checkpoint of the same family.
On standard benchmarks, STT.ai Enhanced achieves around 4.2% Word Error Rate. Real-world accuracy depends on audio quality, accent, and language; for noisy or accented recordings, expect a few percentage points higher WER.
STT.ai Enhanced is a premium model, included with any paid STT.ai plan starting at $5/month. It is not available on the free tier: free and anonymous uploads run Whisper Turbo instead.
STT.ai Enhanced is released under MIT, a permissive open-source license. You can self-host STT.ai Enhanced on your own hardware or use our hosted version — both are commercially usable.
STT.ai Enhanced supports 99 languages. Auto-detection picks the right language for most audio; you can also specify it manually for a small accuracy lift.
STT.ai Enhanced processes audio at about 15.9x real-time on our GPUs. A 1-hour audio file finishes in under 3 minutes; longer files queue and notify by email when done.
STT.ai Enhanced has 1.55B parameters. Larger models tend to be more accurate but slower; STT.ai hosts STT.ai Enhanced on GPU so the parameter count doesn't affect your client-side performance.
STT.ai Enhanced accepts every format STT.ai supports — MP3, WAV, M4A, FLAC, OGG, MP4, MKV, MOV, WebM, AVI, and others. Output as TXT, SRT, VTT, DOCX, JSON, or PDF.
Yes. Speaker diarization runs alongside STT.ai Enhanced for every transcription — each speaker is labeled and you can rename them in the editor afterwards.
Yes. STT.ai Enhanced runs in our private infrastructure — audio is processed and deleted by default. Pro+ adds client-side encryption so transcripts are unreadable without your key, and Private Cloud lets you self-host STT.ai Enhanced entirely in your own VPC.
Use the compare-stt tool to run STT.ai Enhanced against any other supported model on the same audio — you'll see WER, segment count, speaker labels, and confidence scores side-by-side. The STT.ai Enhanced vs Whisper Large V3 comparison is the most commonly run.
Yes, on a paid plan. Specify "stt-ai-enhanced" as the model parameter on the /v1/transcribe endpoint. Keys without an active plan are served Whisper Large v3 Turbo instead. The request still succeeds; it just runs the free model.
Yes. Because STT.ai Enhanced is MIT-licensed, you can self-host it. STT.ai's open-source page lists the project repo and weights. Most production teams use our hosted version to skip GPU procurement, model swaps, and ops.