Saaras v4
Speech-to-text model for 22 Indian languages plus Indian and Global English with transcribe, translate, verbatim, translit and codemix modes.
Specs
- Context
- —
- Max output
- —
- Input price
- INR 30/ hour
- Output price
- —
- Released
- 2026-08-18
Context
- Not published
Pricing · INR / hour
- Input
- INR 30
- Output
- —
- Cache read
- —
- Cache write
- —
Modalities
- Input
- Audio
- Output
- Text
API
- API types
Reasoning
- Not published
Info
- Status
- Active
- Released
- 2026-08-18
- Knowledge cutoff
- —
Features
- Not published
How to call
1 provider
sarvammodel = saaras:v4
audio
POSThttps://api.sarvam.ai/speech-to-text
JSON
Standard format
{
"id": "saaras-v4",
"object": "model",
"created": 1787011200,
"owned_by": "sarvam",
"name": "Saaras v4",
"api": {
"types": [
"audio"
]
},
"limits": {
"context": null,
"input": null,
"output": null
},
"modalities": {
"input": [
"audio"
],
"output": [
"text"
]
},
"reasoning": {
"supported": null,
"efforts": []
},
"pricing": {
"currency": "INR",
"unit": "hour",
"input": 30,
"output": null,
"cache_read": null,
"cache_write": null
},
"features": [],
"info": {
"status": "active",
"release_date": "2026-08-18",
"knowledge_cutoff": null,
"description": "Speech-to-text model for 22 Indian languages plus Indian and Global English with transcribe, translate, verbatim, translit and codemix modes.",
"docs": "https://docs.sarvam.ai/api/getting-started/models/saaras",
"verified_at": "2026-10-11"
}
}Official sources
3