IndiaAI Mission Unveils Indigenous AI Models EKA-1 and Dhīr to Strengthen India's AI Ecosystem
India's IndiaAI Mission has introduced three indigenous foundation models – EKA-1, Dhīr 2B and Dhīr 5B. The models are designed to support multilingual applications, enterprise automation and AI development across Indian languages.

India has taken another step towards strengthening its domestic artificial intelligence ecosystem with the introduction of three indigenous AI foundation models under the IndiaAI Mission.
The newly unveiled models include EKA-1, developed by Soket Labs, and Dhīr 2B and Dhīr 5B, developed by ZenteiQ.
EKA-1 is a 26.3-billion-parameter model trained on approximately 7.5 trillion tokens. It supports all 22 scheduled Indian languages along with English and is designed for applications involving mathematical reasoning, coding and AI agents.
The Dhīr models are multilingual AI assistants supporting English and 10 Indian languages. They are designed for enterprise applications, including document automation, knowledge retrieval and digital services.
An important feature of these models is their potential for secure, on-premise and offline deployment, making them relevant for organisations that require greater control over their data.
The initiative reflects India's broader ambition to develop domestic AI capabilities and reduce dependence on externally developed technologies.









