---
title: "Sarvam AI models: release history and timeline"
description: "Bengaluru company that trains speech, translation, document vision and language models for Indian languages and offers them through its API, open weights and agent products."
type: "release-archive"
canonical_url: "https://theaiengineering.com/releases/sarvam-ai"
markdown_url: "https://theaiengineering.com/releases/sarvam-ai.md"
author: "The AI Engineering Editors"
publisher: "The AI Engineering"
updated: "2026-11-03"
citation: "The AI Engineering Editors. \"Sarvam AI models: release history and timeline.\" The AI Engineering, updated Nov 03, 2026. https://theaiengineering.com/releases/sarvam-ai"
license: "CC BY 4.0"
license_url: "https://creativecommons.org/licenses/by/4.0/"
data_csv: "https://theaiengineering.com/releases.csv"
data_json: "https://theaiengineering.com/releases.json"
---

# Sarvam AI models: release history and timeline

Bengaluru company that trains speech, translation, document vision and language models for Indian languages and offers them through its API, open weights and agent products.

- Headquarters: Bengaluru, India
- Founded: 2023
- Website: https://www.sarvam.ai

21 releases from 1 lab, 9 with open weights. Dates are the labs' official announcement dates.

## November 2026

### Sarvam Prism

Sarvam Prism is a decision model fine-tuned from Sarvam 105B that scores a fixed set of answer options in a single forward pass and returns a probability for each one instead of generating text.

- Date: 2026-11-03
- Lab: Sarvam AI
- Type: LLM
- Weights: Open weights
- Modalities: text → score
- Context window: 131,072 tokens
- Parameters: MoE: 105B total, 10B active
- License: Apache-2.0
- Available at launch: Open weights, API
- Permalink: https://theaiengineering.com/releases/2026/11#sarvam-prism

Highlights:

- Sarvam reported an average of 71.8 across three decision indices at a 44 ms median response on one NVIDIA B200, the top score under 100 ms among the 25 decision models it tested
- Built as a LoRA fine-tune of Sarvam 105B with a small decision head, answering choice, yes/no and ordered-score questions
- Weights were published on Hugging Face under Apache 2.0 in a custom FP8 format that runs on one GPU with about 120 GB of memory
- At launch it was also offered as a hosted model on the Sarvam API; the model card lists a limit of 131,072 tokens per question

Source: [Sarvam Prism: a decision model built on Sarvam 105B | Sarvam AI](https://www.sarvam.ai/blogs/sarvam-prism)

## October 2026

### Bulbul V4 Flash

Bulbul V4 Flash is Sarvam's real-time text-to-speech model, distilled from larger Bulbul V4 research models to keep their expressive delivery in Indian languages at lower latency and cost.

- Date: 2026-10-13
- Lab: Sarvam AI
- Type: Speech
- Weights: Proprietary
- Modalities: text → audio
- Available at launch: API
- Permalink: https://theaiengineering.com/releases/2026/10#bulbul-v4-flash

Highlights:

- Launched with 83 voices across 11 languages and 224 voice and style combinations
- Sarvam reported under 250 ms time to first audio in production
- Priced at launch at ₹20 per 10,000 characters, against ₹30 for Bulbul V3
- Offered at launch through the Sarvam API, Playground and Content Studio; the same post previewed a larger Bulbul V4 Pro model

Source: [Introducing Bulbul V4 Flash text-to-speech for India | Sarvam AI](https://www.sarvam.ai/blogs/bulbul-v4-flash)

## September 2026

### Sarvam Vision 2.1

Sarvam Vision 2.1 is an updated version of Sarvam's document intelligence vision-language model that adds key-value extraction from forms, stronger multi-page table parsing and handwriting recognition for Indian languages.

- Date: 2026-09-24
- Lab: Sarvam AI
- Type: Multimodal
- Weights: Proprietary
- Modalities: text, image → text
- Available at launch: API
- Permalink: https://theaiengineering.com/releases/2026/09#sarvam-vision-2-1

Highlights:

- Sarvam reported an overall olmOCR-Bench score of 87.3 on the official set, the highest in its comparison table
- Post-trained with supervised fine-tuning and reinforcement learning with verifiable rewards on new synthetic and real form and handwriting data
- Kept the earlier design of a vision-language model wrapped by a semantic layout parser and a reading order network
- At launch it was offered through the Document AI Digitise and Extract APIs and the Akshar document agents; Sarvam had previewed these features as Vision 2.0 at its Epoch event

Source: [Sarvam Vision 2.1 document intelligence | Sarvam AI](https://www.sarvam.ai/blogs/sarvam-vision-2-1)

## August 2026

### Saaras V4

Saaras V4 is Sarvam's speech recognition model for 22 Indian languages and English, pairing an audio encoder with a 3B hybrid state-space decoder so one model can return five transcript formats.

- Date: 2026-08-07
- Lab: Sarvam AI
- Type: Speech
- Weights: Proprietary
- Modalities: audio → text
- Available at launch: API
- Permalink: https://theaiengineering.com/releases/2026/08#saaras-v4

Highlights:

- Output modes cover verbatim and normalized transcription, code-mixed text, transliteration and translation
- Added single-pass multi-speaker recognition that combines diarization and transcription
- Sarvam reported the lowest word error rate of the systems it tested on seven English test sets, and a 5.22% language identification error across 22 Indian languages
- Announced at Sarvam Epoch; the changelog added saaras:v4 to the Speech-to-Text API on August 18, 2026, Voice Agents made it generally available on September 2, and a technical post followed on September 25

Source: [Everything we announced at Sarvam Epoch | Sarvam AI](https://www.sarvam.ai/epoch/summary)

### Sarvam 105B Conversations

Sarvam 105B Conversations is a variant of Sarvam 105B post-trained for real-time dialogue, voice agents and multi-turn chat in Indian languages and English.

- Date: 2026-08-07
- Lab: Sarvam AI
- Type: LLM
- Weights: Proprietary
- Modalities: text → text
- Context window: 32,768 tokens
- Available at launch: API
- Permalink: https://theaiengineering.com/releases/2026/08#sarvam-105b-conversations

Highlights:

- On a production benchmark built from Indian customer conversations, Sarvam reported a voice-call score of 82.1, against 74.7 for GPT 5.4 Mini and 78.9 for Gemini 3.5 Flash
- Served under the model ID sarvam-105b-conversations with a 32K-token context window, against 128K for the base model
- Introduced at Sarvam Epoch together with a reasoning-tuned Sarvam 105B update for agent work, both offered through the Sarvam API
- Sarvam's August 18, 2026 changelog priced it the same as sarvam-105b

Source: [Everything we announced at Sarvam Epoch | Sarvam AI](https://www.sarvam.ai/epoch/summary)

## March 2026

### Sarvam 105B

Sarvam 105B is Sarvam's flagship mixture-of-experts reasoning model, trained from scratch in India for reasoning, coding and agentic work with a focus on Indian languages.

- Date: 2026-03-06
- Lab: Sarvam AI
- Type: Reasoning
- Weights: Open weights
- Modalities: text → text
- Context window: 131,072 tokens
- Parameters: MoE: 105B total, 10.3B active
- License: Apache-2.0
- Available at launch: Open weights, API, Consumer app
- Permalink: https://theaiengineering.com/releases/2026/03#sarvam-105b

Highlights:

- About 10.3B active parameters per token, 128 experts and Multi-head Latent Attention; pre-trained on 12T tokens
- Sarvam reported 98.6 on Math500, 71.7 on LiveCodeBench v6 and 90.6 on MMLU
- Weights were released under Apache 2.0 on Hugging Face and AI Kosh, with access through the Sarvam API and the Indus assistant
- Trained on compute provided under the IndiaAI Mission; it had powered the Indus beta that opened on February 20, 2026

Source: [Open-sourcing Sarvam 30B & 105B models | Sarvam AI](https://www.sarvam.ai/blogs/sarvam-30b-105b)

### Sarvam 30B

Sarvam 30B is a mixture-of-experts reasoning model trained from scratch by Sarvam and tuned for fast, low-cost deployment such as multilingual voice calls with tool use.

- Date: 2026-03-06
- Lab: Sarvam AI
- Type: Reasoning
- Weights: Open weights
- Modalities: text → text
- Context window: 65,536 tokens
- Parameters: MoE: 30B total, 2.4B active
- License: Apache-2.0
- Available at launch: Open weights, API
- Retired: 2026-08-18
- Permalink: https://theaiengineering.com/releases/2026/03#sarvam-30b

Highlights:

- 2.4B active parameters per token, 128 experts and grouped-query attention; pre-trained on 16T tokens
- Sarvam reported 88.3 on AIME 25, or 96.7 with tools, and 66.5 on GPQA Diamond
- Weights were released under Apache 2.0 on Hugging Face and AI Kosh, and the model was offered through the Sarvam API
- It already powered Samvaad, Sarvam's conversational agent platform, at release

Sarvam fully deprecated sarvam-30b on its Chat Completions API on August 18, 2026 and pointed users to Sarvam-105B; the open weights stayed on Hugging Face.

Source: [Open-sourcing Sarvam 30B & 105B models | Sarvam AI](https://www.sarvam.ai/blogs/sarvam-30b-105b)

Retirement source: [August 18, 2026 | Sarvam API Docs](https://docs.sarvam.ai/changelog/2026/8/18)

## February 2026

### Sarvam Edge

Sarvam Edge is a set of compact on-device models for speech recognition, speech synthesis and translation in Indian languages, designed to run on phones and laptops without a network connection.

- Date: 2026-02-14
- Lab: Sarvam AI
- Type: Small model
- Weights: Proprietary
- Modalities: text, audio → text, audio
- Parameters: 74M ASR, 24M TTS, about 150M translation
- Permalink: https://theaiengineering.com/releases/2026/02#sarvam-edge

Highlights:

- Speech recognition model with 74M parameters, about 294 MB in FP16, covering 10 Indian languages with automatic language detection
- Text-to-speech model with 24M parameters, about 60 MB, covering 10 Indian languages with 8 multilingual speakers
- Translation model of about 150M parameters covering 11 languages and 110 direct language pairs
- At launch Sarvam said it was building Edge with global device manufacturers and listed no public download or API

Source: [Announcing Sarvam Edge | Sarvam AI](https://www.sarvam.ai/blogs/sarvam-edge)

### Saaras V3

Saaras V3 is Sarvam's speech recognition model for 22 Indian languages and English, built on a new architecture with native streaming for low-latency transcription.

- Date: 2026-02-11
- Lab: Sarvam AI
- Type: Speech
- Weights: Proprietary
- Modalities: audio → text
- Available at launch: API
- Permalink: https://theaiengineering.com/releases/2026/02#saaras-v3

Highlights:

- Sarvam reported a 19.31% word error rate on the 10 most-used languages of the IndicVoices benchmark, against about 22% for Saaras V2.5
- Trained on more than 1 million hours of curated multilingual audio through pre-training, supervised fine-tuning and reinforcement learning
- Decodes incrementally for real-time transcription and also supports batch processing
- At launch it was offered through the Saaras V3 API and a no-code experience in the Sarvam dashboard

Source: [Introducing Saaras V3 speech recognition | Sarvam AI](https://www.sarvam.ai/blogs/asr)

### Bulbul V3

Bulbul V3 is Sarvam's third-generation text-to-speech model, built on an LLM that infers emphasis, pauses and pacing from context to produce natural speech in Indian languages.

- Date: 2026-02-07
- Lab: Sarvam AI
- Type: Speech
- Weights: Proprietary
- Modalities: text → audio
- Available at launch: API
- Permalink: https://theaiengineering.com/releases/2026/02#bulbul-v3

Highlights:

- Launched with more than 35 professional voices across 11 languages, plus voice cloning and a low-latency streaming mode
- In a blind A/B listening study run by Josh Talks across 11 languages, Sarvam reported it was the most preferred model for 8 kHz telephony audio
- Sarvam said language support would later expand to 22 Indian languages
- Sarvam gave all users unlimited API access to the model through February 28, 2026

Source: [Bulbul V3: natural, expressive TTS | Sarvam AI](https://www.sarvam.ai/blogs/bulbul-v3)

### Sarvam Vision

Sarvam Vision is a 3B-parameter state-space vision-language model for document intelligence, OCR and general image understanding in English and 22 Indian languages.

- Date: 2026-02-05
- Lab: Sarvam AI
- Type: Multimodal
- Weights: Proprietary
- Modalities: text, image → text
- Parameters: 3B
- Available at launch: API
- Permalink: https://theaiengineering.com/releases/2026/02#sarvam-vision

Highlights:

- Built by continued pre-training of Sarvam's 3B base model, followed by supervised fine-tuning and reinforcement learning with verifiable rewards
- Handles image captioning, scene text, charts and complex tables, with a layout parser and a reading order network around the model
- Sarvam reported leading results on olmOCR-Bench (English subset) and OmniDocBench v1.5, and introduced a 22-language Sarvam Indic OCR Bench with 20,267 samples
- The Document Intelligence API and dashboard experience were free to use for all of February 2026

Source: [Sarvam Vision multimodal document AI | Sarvam AI](https://www.sarvam.ai/blogs/sarvam-vision)

### Sarvam Audio

Sarvam Audio is a speech recognition model built as an audio extension of Sarvam 3B, a language model trained from scratch on English and 22 Indian languages, that uses conversation context to improve its transcripts.

- Date: 2026-02-03
- Lab: Sarvam AI
- Type: Speech
- Weights: Proprietary
- Modalities: audio, text → text
- Permalink: https://theaiengineering.com/releases/2026/02#sarvam-audio

Highlights:

- Lets applications choose one of five output styles at inference time, from literal transcription to code-mixed, romanized or English-translated text
- Sarvam reported state-of-the-art diarized speech recognition on recordings of 1 to 60 minutes with up to 8 speakers
- Can turn speech directly into a tool call with arguments for voice agents
- At launch Sarvam said it would be available soon on the Sarvam Dashboard

Source: [Sarvam Audio: speech beyond transcription | Sarvam AI](https://www.sarvam.ai/blogs/sarvam-audio)

### Sarvam Dub

Sarvam Dub is a dubbing model that moves speech between Indian languages while keeping the original speaker's voice and fitting the new audio to the video's timing.

- Date: 2026-02-01
- Lab: Sarvam AI
- Type: Speech
- Weights: Proprietary
- Modalities: audio → audio
- Available at launch: Enterprise
- Permalink: https://theaiengineering.com/releases/2026/02#sarvam-dub

Highlights:

- Uses zero-shot voice cloning and builds duration control into speech generation rather than stretching audio afterward
- Sarvam reported state-of-the-art speaker similarity on its benchmark of 64 speakers across 10 Indian languages and English
- Used to dub the monthly Mann Ki Baat address into 11 languages and to dub the Union Budget 2026 speech live into Kannada and Hindi
- Sarvam reported a 6.6x latency reduction over a base PyTorch implementation for live use

Source: [Sarvam Dub: SOTA Indian language dubbing | Sarvam AI](https://www.sarvam.ai/blogs/sarvam-dub)

## June 2025

### Sarvam-Translate

Sarvam-Translate is an open-weights translation model, fine-tuned from Gemma 3 4B with AI4Bharat, that translates whole documents between English and the 22 official Indian languages.

- Date: 2025-06-07
- Lab: Sarvam AI
- Type: LLM
- Weights: Open weights
- Modalities: text → text
- Context window: 8,192 tokens
- Parameters: 4B
- License: GPL-3.0
- Available at launch: Open weights, API
- Permalink: https://theaiengineering.com/releases/2025/06#sarvam-translate

Highlights:

- Handles long-form and structured input such as Markdown, HTML, code and scientific notation while keeping the formatting
- Sarvam reported that language experts rated it well above Gemma3-27B-IT, Llama 4 Scout and Llama-3.1-405B-FP8 in human evaluation
- Weights were released on Hugging Face under GPL-3.0, and the model was offered through Sarvam's API
- The model card lists a maximum context of 8K tokens

Source: [Sarvam Translate for Indian languages | Sarvam AI](https://www.sarvam.ai/blogs/sarvam-translate)

## May 2025

### Sarvam-M

Sarvam-M is a 24B hybrid reasoning model that Sarvam post-trained from Mistral Small with supervised fine-tuning and reinforcement learning to improve Indian language, math and programming performance.

- Date: 2025-05-23
- Lab: Sarvam AI
- Type: Reasoning
- Weights: Open weights
- Modalities: text → text
- Context window: 8,192 tokens
- Parameters: 24B
- License: Apache-2.0
- Available at launch: Open weights, API
- Retired: 2026-06-05
- Permalink: https://theaiengineering.com/releases/2025/05#sarvam-m

Highlights:

- One model with switchable think and non-think modes
- Sarvam reported average gains over the base model of 20% on Indian language benchmarks, 21.6% on math and 17.6% on programming
- Weights were released on Hugging Face under Apache 2.0, with access through Sarvam's playground and API
- Sarvam called its post the first in a series of technical write-ups to accompany planned regular model releases

Sarvam deprecated sarvam-m on its Chat Completions API on June 5, 2026 and directed users to Sarvam-30B and Sarvam-105B; the weights stayed on Hugging Face.

Source: [Sarvam-M multilingual language model | Sarvam AI](https://www.sarvam.ai/blogs/sarvam-m)

Retirement source: [June 5, 2026 | Sarvam API Docs](https://docs.sarvam.ai/changelog/2026/6/5)

## October 2024

### Sarvam-1

Sarvam-1 is a 2B-parameter base language model trained from scratch for 10 Indian languages and English, with a tokenizer that needs far fewer tokens per word for Indic scripts.

- Date: 2024-10-24
- Lab: Sarvam AI
- Type: Small model
- Weights: Open weights
- Modalities: text → text
- Context window: 8,192 tokens
- Parameters: 2B
- License: Sarvam AI Research License
- Available at launch: Open weights
- Permalink: https://theaiengineering.com/releases/2024/10#sarvam-1

Highlights:

- Trained on about 4 trillion tokens, including the 2-trillion-token Sarvam-2T Indic corpus, on 1,024 GPUs of Yotta's Shakti cluster over five days
- Tokenizer fertility of 1.4 to 2.1 tokens per word across the supported languages
- Sarvam reported it outperformed Gemma 2 2B and Llama 3.2 3B on MMLU, ARC-Challenge and IndicGenBench
- Released on Hugging Face as a text-completion model meant for fine-tuning, under a non-commercial research license

Source: [Sarvam 1: the first Indian language LLM | Sarvam AI](https://www.sarvam.ai/blogs/sarvam-1)

## September 2024

### Mayura

Mayura is Sarvam's API translation model for English and Indian languages, trained on formal and colloquial data so its output reads like everyday speech, including code-mixed text.

- Date: 2024-09-06
- Lab: Sarvam AI
- Type: LLM
- Weights: Proprietary
- Modalities: text → text
- Available at launch: API
- Permalink: https://theaiengineering.com/releases/2024/09#sarvam-mayura

Highlights:

- Covered 10 Indian languages at launch and handled respectful second and third person forms
- Offered a gender toggle for first-person translations and kept formatting, URLs, codes and emojis intact
- Could break long multi-clause English sentences, such as legal text, into shorter and more readable Indic sentences
- Sarvam's translation API was part of its August 13, 2024 platform launch; this post introduced the model under the Mayura name

Source: [Making complex info accessible with Mayura | Sarvam AI](https://www.sarvam.ai/blogs/blog-mayura)

## August 2024

### Bulbul v1

Bulbul v1 was Sarvam's first text-to-speech model, a single compact multilingual model built for code-mixed Indian speech.

- Date: 2024-08-23
- Lab: Sarvam AI
- Type: Speech
- Weights: Proprietary
- Modalities: text → audio
- Available at launch: API
- Permalink: https://theaiengineering.com/releases/2024/08#bulbul-v1

Highlights:

- Launched with six voices that keep the same identity across languages, in more than 10 languages
- Trained to pronounce domain terms, local names, numbers and abbreviations common in Indian usage
- Supported pitch and pace control for every voice
- Sarvam's speech synthesis API was part of its August 13, 2024 platform launch; this post named and detailed the model

Source: [Giving voice to India's languages with Bulbul | Sarvam AI](https://www.sarvam.ai/blogs/blog-bulbul)

### Sarvam 2B

Sarvam 2B was Sarvam's first language model trained from scratch, a small base model for 10 Indian languages and English released as an early pre-trained checkpoint.

- Date: 2024-08-13
- Lab: Sarvam AI
- Type: Small model
- Weights: Open weights
- Modalities: text → text
- Parameters: 2B
- License: Sarvam AI Research License
- Available at launch: Open weights
- Permalink: https://theaiengineering.com/releases/2024/08#sarvam-2b

Highlights:

- The Hugging Face card described the v0.5 checkpoint as pre-trained on 2 trillion tokens with no post-training
- Sarvam's tokenizer table showed an average of 2.08 tokens per word across supported languages, against 9.34 for Llama 3.1
- Trained with NVIDIA NeMo on Yotta's Shakti cloud using HGX H100 systems
- Announced at Sarvam's August 13, 2024 platform launch; the fully trained version shipped as Sarvam-1 in October 2024

Source: [Announcing the Sarvam product launch | Sarvam AI](https://www.sarvam.ai/blogs/sarvam-launch)

### Shuka v1

Shuka v1 is an open audio language model that understands spoken questions in Indian languages by joining Sarvam's Saaras v1 audio encoder to Meta's Llama 3 8B Instruct.

- Date: 2024-08-13
- Lab: Sarvam AI
- Type: Speech
- Weights: Open weights
- Modalities: audio, text → text
- License: Llama 3 Community License
- Available at launch: Open weights
- Permalink: https://theaiengineering.com/releases/2024/08#shuka-v1

Highlights:

- The encoder and decoder are linked by a projector of about 60M parameters, the only part that was trained
- Trained on less than 100 hours of audio, with the projector tuned on English and Hindi data
- Sarvam tested it on 11 languages; beyond English and Hindi, it answered spoken questions through zero-shot transfer from the multilingual encoder
- Announced as Shuka 1.0 at Sarvam's August 13, 2024 platform launch and released on Hugging Face

Source: [Announcing the Sarvam product launch | Sarvam AI](https://www.sarvam.ai/blogs/sarvam-launch)

## December 2023

### OpenHathi-7B-Hi-v0.1-Base

OpenHathi-7B-Hi-v0.1-Base was Sarvam's first open model, a 7B Llama 2-based base model extended for Hindi, English and Hinglish.

- Date: 2023-12-12
- Lab: Sarvam AI
- Type: LLM
- Weights: Open weights
- Modalities: text → text
- Parameters: 7B
- License: Llama 2 Community License
- Available at launch: Open weights
- Permalink: https://theaiengineering.com/releases/2023/12#openhathi-7b-hi-v0-1

Highlights:

- First release in Sarvam's OpenHathi series, built with academic partner AI4Bharat
- Added 16K Hindi tokens to the Llama 2 tokenizer for a 48K vocabulary to cut token counts on Hindi text
- Released as a base model meant to be fine-tuned rather than used directly
- Published on Hugging Face under the Llama 2 license

Source: [OpenHathi Series: An Approach To Build Bilingual LLMs Frugally](https://web.archive.org/web/20231217000713/https://www.sarvam.ai/blog/announcing-openhathi-series)

## Cite and reuse this data

These release records are free to reuse under CC BY 4.0, including commercially. Credit The AI Engineering and link to this page's URL. Dates are the labs' official announcement dates, and every record links to its source. List fields in the CSV are separated by " | ".

The AI Engineering Editors. "Sarvam AI models: release history and timeline." The AI Engineering, updated Nov 03, 2026. https://theaiengineering.com/releases/sarvam-ai

HTML link:

```html
<a href="https://theaiengineering.com/releases/sarvam-ai">Sarvam AI models: release history and timeline</a> (The AI Engineering)
```

- Title: Sarvam AI models: release history and timeline
- Author: The AI Engineering Editors
- Publisher: The AI Engineering
- Updated: 2026-11-03
- URL: https://theaiengineering.com/releases/sarvam-ai
- Markdown: https://theaiengineering.com/releases/sarvam-ai.md
- License: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/)
- Full dataset as CSV: https://theaiengineering.com/releases.csv
- Full dataset as JSON: https://theaiengineering.com/releases.json
