Indus
Tech Mahindra's Hindi-first sovereign LLM built under Project Indus.
About
Indus is Tech Mahindra's sovereign, Indic-first large language model, developed by its Makers Lab research hub under the Project Indus initiative. It started as a 1.2B Hindi-first foundation model built from scratch with authentic Indian linguistic and cultural context — designed to converse in Hindi and 37+ dialects — and was scaled to an 8B education-focused LLM launched at the India AI Impact Summit in February 2026.
The current generation is built on NVIDIA's stack (NeMo framework, NIM microservices), with nearly half a billion synthetic tokens generated via NeMo Data Designer to counter Hindi data scarcity, and supports agentic AI workflows so developers can build Hindi-fluent autonomous agents. A telecom-sector multimodal variant launched in March 2025, and Indus 2.0 extends the model to retail, banking, healthcare, and citizen services.
Honest caveats: Indus weights are not open-sourced — this is a sovereign commercial program, not an open release. Performance and adoption figures are company-reported, and the education LLM's real-world classroom impact is still early.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 7/10
Scored 7/10 (metrics tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 76; Trending score 70/100 from researcher assessment.