Sarvam's Saaras V4 Brings All 22 Indian Languages Into One ASR Model

Sarvam's new speech-to-text model beats Deepgram Nova-3 and GPT-4o Transcribe on noisy audio while covering all 22 Indian languages.

Source

Sarvam

Published

Sep 23, 2026, 11:47 AM

Author

AlphaSignal Newsroom

Read

1 min read

Sarvam's new speech-to-text model beats Deepgram Nova-3 and GPT-4o Transcribe on noisy audio while covering all 22 Indian languages.

Reporting is indexed from AlphaSignal. Rights remain with the original publisher and cited sources.

Next reads