Timothyves: new new stub

2026-05-01T14:49:06Z

new new stub

New page

{{Short description|Indian artificial intelligence company}}
{{Infobox company
| name = Sarvam AI
| logo = Sarvam AI logo.svg
| type = [[Privately held company|Private]]
| industry = [[Artificial intelligence]]
| founded = {{start date and age|2023|08}}
| founders = Vivek Raghavan; Pratyush Kumar
| headquarters = [[Bengaluru]], [[Karnataka]], India
| area_served = [[India]]
| products = [[Large language model]]s; Speech recognition systems; [[Vision-language model]]s
| website = https://www.sarvam.ai/ (Main Website)
* https://indus.sarvam.ai/ (AI Agent)
}}

'''Sarvam AI''' is an [[List of companies of India|Indian]] [[artificial intelligence]] company headquartered in [[Bengaluru]], [[Karnataka]]. Founded in 2023, the company develops [[large language models]] (LLMs) and multimodal AI systems with a focus on [[Indo-Aryan languages|Indian languages]] and region-specific use cases.<ref name="TechCrunch2023">{{cite news |last=Singh |first=Manish |title=Five-month-old Indian AI startup Sarvam scores $41M funding |url=https://techcrunch.com/2023/12/06/indias-sarvam-ai-raises-41-million-from-lightspeed-khosla-peak-xv/ |work=TechCrunch |date=6 December 2023}}</ref>

The company has received venture capital backing and has participated in [[Government of India|government]]-supported AI initiatives, including India's sovereign large language model programme under the IndiaAI Mission.<ref name="IndianExpress2025">{{cite news |title=Sarvam AI to build India's first sovereign LLM with reasoning and voice capabilities |url=https://indianexpress.com/article/technology/artificial-intelligence/sarvam-ai-to-build-indias-first-sovereign-llm-with-reasoning-and-voice-capabilities-9967108/ |work=The Indian Express |date=27 April 2025}}</ref>

==History==

Sarvam AI was founded in August 2023 by Vivek Raghavan and Pratyush Kumar, who were previously associated with AI4Bharat at the [[IIT Madras|Indian Institute of Technology Madras]].<ref name="TechCrunch2023"/> In December 2023, the company announced a combined seed and Series A funding round of approximately US$41 million. The round was led by [[Lightspeed Venture Partners]], with participation from Peak XV Partners and [[Khosla Ventures]].<ref name="MintFunding2023">{{cite news |title=AI startup Sarvam AI raises $41 million from Lightspeed, Khosla Ventures |url=https://www.livemint.com/ai/ai-startup-sarvam-ai-raises-41-million-from-lightspeed-khosla-ventures-11701859187313.html |work=Mint |date=6 December 2023}}</ref>

In April 2025, the [[Ministry of Electronics and Information Technology]] (MeitY) selected Sarvam AI as one of the companies to develop an indigenous foundational model under the IndiaAI Mission.<ref name="Mint2025">{{cite news |last=Gupta |first=Aman |title=MeitY picks Sarvam AI to lead India's sovereign LLM effort; Vaishnaw vows global competition |url=https://www.livemint.com/technology/tech-news/india-taps-sarvam-ai-to-build-first-sovereign-llm-under-indiaai-mission-ashwini-vaishnaw-11745660295745.html |work=Mint |date=26 April 2025}}</ref> As part of the initiative, the company received access to government-supported computing infrastructure, including GPUs allocated for model training over a specified period.<ref name="Mint2025"/>

In February 2026, Sarvam AI introduced two large language models at the [[India AI Impact Summit 2026|AI Impact Summit]] held at [[Bharat Mandapam]], [[New Delhi]].<ref name="FortuneIndia2026">{{cite news |title=Sarvam AI launches 30B and 105B models tailored for India-focused deployment |url=https://www.fortuneindia.com/business-news/sarvam-ai-launches-30b-and-105b-models-tailored-for-india-focused-deployment/130517 |work=Fortune India |date=18 February 2026}}</ref>

== Products and technology ==

Sarvam AI develops language models trained on datasets that include multiple [[Languages with official recognition in India|Indian languages]] and code-mixed text. The company uses mixture-of-experts (MoE) architectures in some of its models.<ref name="FortuneIndia2026"/>

=== Foundational language models ===

On 18 February 2026, the company announced the release of two foundational models:

* Sarvam-30B – A 30-billion parameter model based on a mixture-of-experts design. According to company disclosures reported by the media, the model activates approximately 1 billion parameters per token and supports a 32,000-token context window.<ref name="FortuneIndia2026"/>
* Sarvam-105B – A 105-billion parameter model activating approximately 9 billion parameters per token, with a 128,000-token context window. The model is positioned for complex reasoning and enterprise applications.<ref name="TOI2026">{{cite news |last=Vaitheeswaran |first=B. |title=Sarvam rolls out 105-bn parameter AI LLM model |url=https://timesofindia.indiatimes.com/business/india-business/sarvam-rolls-out-105-bn-parameter-ai-llm-model/articleshow/128532374.cms |work=The Times of India |date=19 February 2026}}</ref>
On 20th February 2026, the company released a beta version of the Sarvam-105B model which is named Indus. It is available on the Apple App Store, Google Play Store and the web.<ref>{{Cite web |title=Indus AI app: Sarvam launches desi ChatGPT rival on app stores |url=https://www.deccanherald.com/technology/artificial-intelligence/sarvam-brings-chatgpt-challenger-indus-aiapp-available-on-google-play-apple-app-store-3905567 |access-date=2026-02-20 |website=Deccan Herald |language=en}}</ref>

=== Speech and vision systems ===

Sarvam AI has also developed multimodal systems including [[Speech recognition|speech-to-text]] and [[vision-language model]]s. Its speech model, referred to as Saaras V3 in company materials, supports multiple Indian languages.<ref name="SarvamModelsOfficial">{{cite web |title=Sarvam Models Overview |url=https://www.sarvam.ai/models/ |publisher=Sarvam AI |access-date=20 February 2026}}</ref>

The company has also introduced a vision-language model known as Sarvam Vision, intended for document understanding and [[optical character recognition]] (OCR) in [[Indian script]]s.<ref name="BISResearch2026">{{cite news |title=India’s Sarvam AI Emerges in Sovereign AI Race |url=https://bisresearch.com/news/indias-sarvam-ai-emerges-as-a-sovereign-powerhouse-outperforming-chatgpt-and-google-gemini |work=BIS Research |date=10 February 2026}}</ref>

=== Devices ===
'Sarvam Kaze' is an indigenous AI-powered wearable glass that listens, understands, and captures what users see the world through their eyes in real time. The device supports more than 10 Indian languages, enabling voice-based interaction and potentially real-time translation. The company plans to launch the device in May 2026.<ref>{{Cite web |title=What is Sarvam Kaze? India's first indigenous glasses PM Modi wore at AI Impact Summit 2026 |url=https://www.wionews.com/trending/what-is-sarvam-kaze-india-s-first-indigenous-glasses-pm-modi-wore-at-ai-impact-summit-2026-1771319066279 |access-date=2026-02-20 |website=Wion |language=en}}</ref>
== Startup support ==
In March 2026, Sarvam AI launched the Sarvam Startup Program, an initiative providing selected early-stage companies with 6–12 months of API credits scaled to their needs, priority engineering support, and access to production infrastructure for developing multilingual AI applications in areas such as speech, translation, and large language models.<ref>{{cite web |url=https://www.cnbctv18.com/business/startup/sarvam-ai-launches-startup-programme-heres-all-about-it-and-what-it-offers-ws-l-19863357.htm |title=Sarvam AI launches startup programme: Here's all about it and what it offers |website=CNBC TV18 |date=6 March 2026 |access-date=7 March 2026}}</ref><ref>{{cite web |url=https://analyticsindiamag.com/ai-news/sarvam-ai-launches-startup-programme-to-support-early-stage-companies |title=Sarvam AI Launches Startup Programme To Support Early-Stage Companies (via LinkedIn post reference, but use full article if available) |website=Analytics India Magazine |date=March 2026 |access-date=7 March 2026}}</ref>
==Open-source release==
In February 2026, Sarvam AI announced and open-sourced two large language models: Sarvam 30B (30 billion parameters) and Sarvam 105B (105 billion parameters, using a Mixture-of-Experts architecture with 10.3 billion active parameters). Both models were trained from scratch on datasets focused on Indian languages and support advanced reasoning, multilingual tasks, mathematics, and coding.<ref>{{cite web |url=https://techcrunch.com/2026/02/18/indian-ai-lab-sarvams-new-models-are-a-major-bet-on-the-viability-of-open-source-ai/ |title=Indian AI lab Sarvam's new models are a major bet on the viability of open source AI |website=TechCrunch |date=18 February 2026 |access-date=7 March 2026}}</ref><ref>{{cite web |url=https://www.sarvam.ai/blogs/sarvam-30b-105b |title=Open-Sourcing Sarvam 30B and 105B |website=Sarvam AI |date=6 March 2026 |access-date=7 March 2026}}</ref>
The models are hosted on Hugging Face under the Apache License and are intended for enterprise and developer applications in Indian languages.<ref>{{cite web |url=https://huggingface.co/sarvamai/sarvam-105b |title=sarvamai/sarvam-105b |website=Hugging Face |access-date=7 March 2026}}</ref>

The models were subsequently released as open source under the Apache License 2.0, with model weights made available on Hugging Face (sarvamai/sarvam-30b and sarvamai/sarvam-105b) and AIKosh in early March 2026.<ref>{{cite web |url=https://www.sarvam.ai/blogs/sarvam-30b-105b |title=Open-Sourcing Sarvam 30B and 105B |website=Sarvam AI |date=6 March 2026 |access-date=7 March 2026}}</ref><ref>{{cite web |url=https://huggingface.co/sarvamai/sarvam-105b |title=sarvamai/sarvam-105b |website=Hugging Face |access-date=7 March 2026}}</ref><ref>{{cite news |url=https://timesofindia.indiatimes.com/technology/tech-news/sarvam-takes-on-google-openai-and-anthropic-launches-105-billion-parameter-open-source-model-for-india/articleshow/128509486.cms |title=Sarvam takes on Google, OpenAI and Anthropic; launches 105-billion parameter open-source model for India |website=The Times of India |date=18 February 2026 |access-date=7 March 2026}}</ref>

== Government and institutional collaborations ==

In 2025, Sarvam AI was selected to contribute to India's sovereign AI model initiative under the IndiaAI Mission.<ref name="IndianExpress2025"/> The initiative aims to support domestic AI infrastructure and model development.

In March 2025, the [[Unique Identification Authority of India]] (UIDAI) announced a collaboration with Sarvam AI to integrate AI-based voice interactions and multilingual support into [[Aadhaar]]-related services.<ref name="UIDAI2025">{{cite news |last=Vishwakarma |first=Sumit |title=UIDAI partners with Bengaluru-based Sarvam AI to enhance user experience of Aadhaar services |url=https://indianstartupnews.com/news/uidai-partners-with-sarvam-ai-to-enhance-user-experience-of-aadhaar-services-8868627 |work=Indian Startup News |date=19 March 2025}}</ref>

Sarvam AI has also worked with AI4Bharat and academic institutions on language datasets and speech research projects.<ref name="IndiaAIIndicVoices">{{cite web |title=IIT Madras, AI4Bharat, and Sarvam AI launch IndicVoices |url=https://indiaai.gov.in/article/iit-madras-ai4bharat-and-sarvam-ai-launch-indicvoices-a-milestone-in-indian-speech-recognition |publisher=IndiaAI |date=28 November 2024}}</ref>

== Industry participation ==

Sarvam AI presented its foundational models at the India [[India AI Impact Summit 2026|AI Impact Summit 2026]] in [[New Delhi]].<ref name="FortuneIndia2026"/> The company has also been listed among Indian members of the AI Alliance, a consortium focused on open-source artificial intelligence initiatives.<ref name="BusinessStandard2024">{{cite news |title=AI Alliance adds 7 Indian members including Infosys, AI4Bharat & Sarvam AI |url=https://www.business-standard.com/companies/news/ai-alliance-adds-7-indian-members-including-infosys-ai4bharat-sarvam-ai-124091800979_1.html |work=Business Standard |date=18 September 2024}}</ref>

== List of models ==

{| class="wikitable sortable"
|+ List of Sarvam AI models
|-
! Name
! Release date
! License
! {{Ref.}}
|-
| Sarvam-1
| {{dts|2024-10}}
| {{partial|Sarvam AI Research}}
| <ref>{{cite web |last1=Jindal |first1=Siddharth |title=Sarvam AI Launches Sarvam-1, Outperforms Gemma-2 and Llama-3.2 |url=https://analyticsindiamag.com/ai-news-updates/sarvam-ai-launches-sarvam-1-outperforms-gemma-2-and-llama-3-2/ |website=Analytics India Magazine |access-date=20 April 2026 |archive-url=https://web.archive.org/web/20250725131430/https://analyticsindiamag.com/ai-news-updates/sarvam-ai-launches-sarvam-1-outperforms-gemma-2-and-llama-3-2/ |archive-date=25 July 2025 |language=en |date=24 October 2024 |url-access=limited}}</ref><ref>{{cite web |title=LICENSE.md · sarvamai/sarvam-1 |url=https://huggingface.co/sarvamai/sarvam-1/blob/d3880226af5d8adffd44250463f31ae6fe16073b/LICENSE.md |via=[[Hugging Face]] |access-date=20 April 2026 |date=23 October 2024}}</ref>
|-
| Sarvam-M
| {{dts|2025-5}}
| rowspan=2 {{yes|[[Apache 2.0]]}}
| <ref>{{cite web |title=Sarvam AI debuts flagship open-source LLM with 24 billion parameters |url=https://indianexpress.com/article/technology/artificial-intelligence/sarvam-ai-debuts-flagship-open-source-llm-with-24-billion-parameters-10026071/ |website=[[The Indian Express]] |access-date=20 April 2026 |date=25 May 2025}}</ref><ref>{{cite web |last1=Jindal |first1=Siddharth |title=India Loves Llama, But Flirts with Mistral and Qwen |url=https://analyticsindiamag.com/ai-startups/india-loves-llama-but-flirts-with-mistral-and-qwen |website=Analytics India Magazine |access-date=20 April 2026 |archive-url=https://web.archive.org/web/20250527131459/https://analyticsindiamag.com/ai-startups/india-loves-llama-but-flirts-with-mistral-and-qwen/ |archive-date=27 May 2025 |date=27 May 2025 |url-access=limited}}</ref><ref>{{cite web |title=LICENSE.txt · sarvamai/sarvam-m |url=https://huggingface.co/sarvamai/sarvam-m/blob/21303daf2ee89055dc5120484be9ea0f4ac397d1/LICENSE.txt |via=[[Hugging Face]] |access-date=20 April 2026 |date=23 May 2025}}</ref>
|-
| Sarvam-105B and Sarvam-30B
| {{dts|2026-2}}
| <ref>{{cite web |last1=Das |first1=Avik |title=AI startup Sarvam launches two made-in-India large language models |url=https://www.business-standard.com/technology/tech-news/ai-startup-sarvam-launches-two-made-in-india-large-language-models-126021801397_1.html |website=[[Business Standard]] |access-date=20 April 2026 |date=18 February 2026}}</ref><ref>{{cite web |title=Sarvam releases open-weight models debuted at AI Summit: How they compare with DeepSeek, Gemini |url=https://indianexpress.com/article/technology/artificial-intelligence/sarvam-releases-open-weight-models-ai-summit-deepseek-gemini-10571340/ |website=[[The Indian Express]] |access-date=20 April 2026 |date=10 March 2026}}</ref>
|}

== See also ==

{{Portal|Free and open-source software|Business|India}}

* [[List of large language models]]
* [[Lists of open-source artificial intelligence software]]
* [[Generative artificial intelligence]]
* [[Large language model]]
* [[Artificial intelligence in India]]

{{clear}}

== References ==
{{Reflist}}

==External links==
*{{url|www.sarvam.ai|Official site}}
*[https://indus.sarvam.ai/ AI Chatbot]
*[https://huggingface.co/sarvamai Sarvam AI] on [[Hugging Face]]

[[Category:Artificial intelligence companies]]
[[Category:Companies based in Bengaluru]]
[[Category:Artificial intelligence laboratories]]
[[Category:2023 establishments in India]]
[[Category:Open-source artificial intelligence]]

Sarvam AI - Revision history

Timothyves: new new stub