Skip to content
Moojaz — موجز
Search
Menu
HomeRegionsMaghrebAfricaMiddle EastEuropeAsiaAmericasWorldSectionsPoliticsEconomySecurityHealthScienceTechnologyAIEnvironmentCultureSportsSearch
India
File

India: EKA-1 and Dhīr AI models

Published

This dossier tracks IndiaAI-backed EKA-1 from Soket Labs and Dhīr from ZenteiQ, their language coverage and evidence of practical use.

Latest event : 1 key moments
How we use AIReport an error
Official presentation of EKA-1 and Dhīr models in India.
Presentation in New Delhi on 3 November 2026. Press Information Bureau, Government of India. — Press Information Bureau, Government of India; reuse permitted under PIB copyright policy https://www.pib.gov.in/content/3604_2_CopyrightPolicy.aspx?lang=1&reg=6

Where things stand

As of 3 November 2026, India’s IT ministry has presented EKA-1, Dhīr 2B and Dhīr 5B. Announced capabilities, including local Dhīr use, should be distinguished from verified comparative performance.

Browse the full timeline

Reported information

India presents three AI models, including offline-capable Dhīr

India’s Ministry of Electronics and IT presented Soket Labs’ EKA-1 and ZenteiQ’s Dhīr 2B and 5B in New Delhi on 3 November 2026. The models received support from the public IndiaAI programme.

Official presentation of EKA-1 and Dhīr models in India.
Presentation in New Delhi on 3 November 2026. Press Information Bureau, Government of India. — Press Information Bureau, Government of India; reuse permitted under PIB copyright policy https://www.pib.gov.in/content/3604_2_CopyrightPolicy.aspx?lang=1&reg=6

The ministry says EKA-1 has 26.3 billion parameters and was trained on roughly 7.5 trillion tokens, the text units a model processes. It targets maths, coding and reasoning across all 22 constitutionally scheduled Indian languages and English.

Dhīr models are announced as open source, supporting English and ten Indian languages, with local execution without an internet connection. Developers target sensitive on-premise applications among other uses. The consulted sources provide no independent comparison establishing their performance.

Sources 3