Now Reading: MeitY unveils three indigenous AI models under IndiaAI Mission
-
01
MeitY unveils three indigenous AI models under IndiaAI Mission
The Story in Brief
The Ministry of Electronics and Information Technology on Tuesday, 3 November, unveiled three indigenous AI foundation models developed under the IndiaAI Mission: EKA-1 by Soket Labs and Dhīr 2B and Dhīr 5B by ZenteiQ AiTech. The models are designed for secure, offline deployment in sectors such as banking, cybersecurity, defence and education. EKA-1, a 26.3-billion-parameter model trained on 7.5 trillion tokens, supports all 22 scheduled Indian languages and English. The Dhīr models support English and 10 Indian languages and can run on mobile phones without internet.
MeitY Secretary S. Krishnan said the initiative strengthens India’s strategic autonomy in AI and expands the availability of models suited to domestic requirements. Both companies received subsidised national AI compute infrastructure. The models are accessible through the AIKosh platform for researchers, enterprises and the public. No benchmark results, deployment details or pricing were disclosed.
Indian Opinion Analysis
All four outlets report the same event in near-identical wire-like terms, emphasising the launch of three indigenous AI models under the IndiaAI Mission. Livemint, Hindustan Times, Rediff and Inc42 all lead with the MeitY unveiling, name the same models, developers and secretary S. Krishnan, and highlight air-gapped deployment and Indian-language support. There is no divergence in framing: no outlet omits a significant fact, and none adds an editorial slant. The coverage is uniform straight reporting, leaving the story's sole point of tension, benchmark performance, pricing and rollout timelines, untouched in every version. The next concrete indicator is whether the models' actual adoption, especially in banking and defence, meets the government's stated ambition.
Coverage: 4 sources, 4 neutral
Sources (4): livemint.com (neutral report), hindustantimes.com (neutral report), rediff.com (neutral report), inc42.com (neutral report)
This brief was synthesised by AI from the 4 sources linked above, so one read covers every framing they carry. Methodology and corrections.
