India unveiled three homegrown foundation models—EKA-1, Dhir 2B, and Dhir 5B—developed by Soket Labs and ZenteiQ. *(Nov. 3)

On November 3, 2026, MeitY Secretary S. Krishnan unveiled three IndiaAI Mission foundation models in New Delhi: EKA-1, built by Soket Labs, and Dhīr 2B and Dhīr 5B, from ZenteiQ AiTech Innovations. EKA-1 has 26.3 billion parameters and was trained on about 7.5 trillion tokens; it supports all 22 scheduled Indian languages plus English and targets mathematics, coding, reasoning and agentic tasks. Dhīr’s two open-source models support English and 10 Indian languages, with secure, on-premise, offline-capable use for document automation and information retrieval. ZenteiQ says they can run on GPUs or TPUs and their lightweight design enables mobile deployment. The models address a practical gap: organisations in government, defence, education and industry can process sensitive information locally or serve users with weak connectivity, rather than rely on cloud access. Both companies were selected under the Mission’s Foundation Models pillar and received subsidised national compute and research support.