- Sarvam AI has launched the Epoch Builder Edition, a platform designed to help organizations build, fine-tune, and deploy large language models (LLMs) tailored for Indian languages and use cases.
- The platform will be available in private preview from August 2026, with general availability expected in the fourth quarter of 2026.
- Sarvam AI's platform features 7-billion and 70-billion parameter multilingual LLMs, trained on 2 trillion tokens and supporting over 10 Indian languages.
- Sarvam AI has partnered with IBM to develop a sovereign AI stack for governments and regulated sectors in India, utilizing IBM's GovTech AI Innovation Centre in Lucknow.
- The collaboration aims to pilot AI solutions for citizen services, grievance redressal, document processing, and administrative workflows.
- This week, discussions around India's sovereign AI ambitions have intensified, highlighting the country's focus on developing indigenous AI capabilities.
- The Epoch Builder Edition platform provides access to base models, training infrastructure, datasets, and deployment tools, aiming to reduce reliance on foreign AI models.
- Sarvam AI's platform includes tools for red-teaming, toxicity checks, and benchmarking tailored to India's cultural and linguistic context.
- The company plans to build a trillion-plus-parameter foundation model from scratch in India, which is seen as essential for competing globally in various sectors.
Sarvam AI has unveiled its Epoch Builder Edition, a platform designed to empower organizations in India to create and deploy large language models (LLMs) tailored for local languages and contexts. This initiative aims to diminish reliance on foreign AI technologies, enhancing India's self-sufficiency in artificial intelligence.1
The platform features 7-billion and 70-billion parameter multilingual LLMs, trained on 2 trillion tokens and supporting over 10 Indian languages including Hindi, Kannada, Tamil, Marathi, and Bengali. It offers developers and enterprises access to GPU clusters for distributed training, reinforcement learning from human feedback (RLHF), and tools for red-teaming and toxicity checks tailored to India's cultural context.38
Sarvam AI's collaboration with IBM aims to create a sovereign AI stack, utilizing IBM's GovTech AI Innovation Centre in Lucknow for pilot projects in citizen services and administrative workflows. The partnership will focus on developing AI solutions for government and regulated sectors, enhancing public service delivery.45
Co-founder Pratyush Kumar emphasized the importance of building indigenous AI capabilities, stating, "It is important that we have our own ability to build models that we need for a variety of applications..." The platform is set for private preview in August 2026, with general availability expected in the fourth quarter of the same year.2
Sarvam AI's ambitious plans include developing a trillion-plus-parameter foundation model from scratch, aiming to compete globally in sectors like coding and cybersecurity.9
The launch of Epoch Builder Edition marks a significant step towards establishing a comprehensive AI ecosystem in India, with a focus on local languages and applications.
“Epoch Builder Edition features 7-billion and 70-billion parameter multilingual models trained on 2 trillion tokens and supporting more than 10 Indian languages, with general availability planned for the fourth quarter of 2026. Sarvam AI also announced partnerships with three IITs and two state governments for pilot projects in education and public service delivery.”
Forbes India
