Models/Sarvam AI/Saaras v3

Saaras v3

Sarvam AI 📝 Transcription

Sarvam's speech-to-text with built-in translation — transcribes Indian language audio and translates to English in one step. Domain-aware processing for improved accuracy.

Specifications

Speed●●●●● Fast
ModalitiesInput: audio  ·  Output: text
FeaturesTools: No Streaming: No

Pricing

Included with plan

Capabilities

translation
Use Saaras v3 on Zubnet →

More from Sarvam AI

S
Saarika v2.5
Sarvam's speech-to-text model — accurate transcription for 22 Indian languages with optional speaker diarization. REST, WebSocket streaming, and batch processin
📝 Transcription
S
Sarvam-M
Sarvam's 24B multilingual LLM — fluent in all 22 official Indian languages + English. OpenAI-compatible API. Free tier with no per-token charges. Supports wiki
🧠 Language Model
S
Sarvam Translate v1
High-quality machine translation across 22 Indian languages and English. Handles complex sentence structures and idiomatic expressions with cultural context awa
🧠 Language Model
M
Mayura v1
Sarvam's multilingual translation model — translates between English and 22 Indian languages with high accuracy. Named after the Indian peacock.
🧠 Language Model