Audio / Speech-to-Text

Saaras V3

Speech-to-text model from Sarvam AI · released Feb 2026 · vendor-reported (not yet ranked on an independent leaderboard)

Unrankednot on an independent leaderboard yet
Feb 2026release

Specifications

Languages
23 (22 Indian + English)
Output modes
transcribe, translate, verbatim, translit, codemix
API model id
saaras:v3

Notes

  • Not ranked on the Artificial Analysis speech-to-text leaderboard at the time of writing.

How this is measured

AA-WER v2 (Speech-to-Text)

Word error rate, % of words transcribed incorrectly (lower is better)

Artificial Analysis Word Error Rate index for non-streaming transcription. AA-WER v2 combines three datasets: AA-AgentTalk (50%), VoxPopuli-Cleaned-AA (25%) and Earnings22-Cleaned-AA (25%). Measured independently on dedicated hardware.

Release & sources

Release
Feb 2026 (month precision)
Release source
Sarvam changelog (speech-to-text endpoints migrated to saaras:v3, February 2026)

More STT models from Sarvam AI