Sarvam AISaaras
Saaras V2.5
Saaras V2.5 is the older speech-to-English model. Sarvam says it continues to work during transition, but is being phased out.
Deprecated API; check access details
Migration changes the endpoint
V2.5 uses speech-to-text-translate. Newer Saaras models use speech-to-text with the translate mode. The notice gives no fixed removal date.Saaras model guide
Practical use
Practical guidance. Compare translated meaning and named entities before switching endpoints. Preserve the original audio so a change in translation can be checked against the source rather than against an older model's output.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Status | Deprecated API; check access details |
| Access | API |
| Model identifiers | saaras:v2.5 |
Related models
A speech-recognition model for Indian languages and global English with normalized, verbatim, transliterated, code-mixed and translated output.
Launch date not confirmedAPI