Skip to main content
The speech_models parameter lets you specify which model(s) to use for transcription. For new accounts created on or after November 17, 2026, leaving out speech_models defaults to ["assemblyai-stt-3-7"]. For existing accounts, see the assemblyai-stt-3-7 migration guide.

Available models

Universal-3.5 Pro (universal-3-5-pro) and Universal-3 Pro (universal-3-pro) are deprecated. The retirement date will be announced. After retirement, if no action is taken, your requests to these models will automatically route to assemblyai-stt-3-7. See Model versions and deprecations.If you would like to migrate to the latest model now, see the assemblyai-stt-3-7 migration guide.

Set the model

You can change the model by setting the speech_models in the POST request body:

Transcribe all 99 languages by combining speech models

When you set speech_models to ["assemblyai-stt-3-7"], whether AssemblyAI falls back to another model depends on how you set the language:
  • language_detection: true: AssemblyAI always transcribes with AssemblyAI-STT-3.7 and never falls back to another model. To fall back to Universal-2 for detected languages that AssemblyAI-STT-3.7 doesn’t support, set speech_models to ["assemblyai-stt-3-7", "universal-2"].
  • language_code: If you set a language code that AssemblyAI-STT-3.7 doesn’t support, AssemblyAI falls back to Universal-2, even if universal-2 isn’t in the list.

AssemblyAI-STT-3.7 takes priority in the list

If you list multiple models alongside assemblyai-stt-3-7 via speech_models, AssemblyAI uses AssemblyAI-STT-3.7 for every language it supports, regardless of its position in the list. Other models in the list are used only for languages AssemblyAI-STT-3.7 doesn’t support. This applies whether you set language_code or use language_detection. If speech_models is set and doesn’t include assemblyai-stt-3-7, AssemblyAI uses the first model in the list, from left to right, that supports the language. This behavior hasn’t changed. See Supported languages for the languages each model supports.

Identify the model used

After transcription completes, you can check which model was actually used to process your request by reading the speech_model_used field. This is useful because the model used may differ from the first model in your list. AssemblyAI-STT-3.7 takes priority when it’s in the list, and the system may fall back to a different model depending on language support.

Complete example

Here is the full working code that demonstrates model selection with error handling: