Speech-to-Text REST API models

Available speech-to-text models on the Telnyx STT REST API. Compare general, telephony, and specialty models by language, accuracy, and latency.

Your choice of model determines which audio formats are accepted, what language values are valid, and what response fields are available.

	openai/whisper-large-v3-turbo	openai/whisper-tiny	deepgram/nova-3	deepgram/nova-3-medical
Default	Yes
Audio formats	All 10	All 10	mp3, wav only	mp3, wav only
Language	80+ languages, auto-detected	50+ languages, auto-detected	~49 languages plus multi	English only (en and regional variants such as en-US, en-GB)
Timestamps	No	No	Word-level (via model_config)	Word-level (via model_config)
Diarization	No	No	Yes (via model_config)	Yes (via model_config)
Smart formatting	No	No	Yes (via model_config)	Yes (via model_config)
model_config	Returns 400	Returns 400	Deepgram pass-through	Deepgram pass-through

deepgram/nova-2 and deepgram/nova-2-medical are also accepted and behave the same way, with a smaller language set (~33 languages plus multi for nova-2). Prefer the nova-3 models unless you have a specific reason.

openai/whisper-large-v3-turbo

Default model. Multilingual. Auto-detected if language omitted. See Whisper docs for the full language list. Returns text only — no timestamps regardless of response_format.

openai/whisper-tiny

Lightweight, lowest resource usage. Multilingual (50+ languages, auto-detected). Returns text only — no timestamps.

deepgram/nova-3

Highest accuracy for English. Advanced features (diarization, word timestamps, smart formatting, numerals, punctuation) available via model_config. Defaults language to en if omitted. Can also set language inside model_config — top-level field takes precedence. See Deepgram language docs for details.

deepgram/nova-3-medical

Nova-3 tuned for clinical speech — medication names, dosages, and medical terminology. Same feature set and model_config pass-through as deepgram/nova-3.

English only. language accepts en and its regional variants (en-US, en-GB, en-AU, en-IN, en-NZ); any other value, including multi, returns a 400 with language_not_found.

curl -X POST https://api.telnyx.com/v2/ai/audio/transcriptions \
  -H "Authorization: Bearer $TELNYX_API_KEY" \
  -F "model=deepgram/nova-3-medical" \
  -F "language=en" \
  -F "response_format=verbose_json" \
  -F 'model_config={"diarize":true,"smart_format":true,"punctuate":true}' \
  -F "file=@consultation.wav"

deepgram/nova-2 and deepgram/nova-2-medical

Previous generation. Same request shape and model_config pass-through as the nova-3 models; deepgram/nova-2 covers ~33 languages plus multi, and deepgram/nova-2-medical is English only. Prefer nova-3 unless you have a specific reason.