Latest / Frontier models and safety
NVIDIA and IIT Madras announce 2.8 trillion-token Indic language corpus
See how this story connectsNVIDIA and IIT Madras’s AI4Bharat lab announced Nemotron-pretraining-Indic-v1, a corpus of approximately 2.8 trillion tokens across the 22 languages listed in India’s Constitution, with release due soon. It has native-script data for all 22 languages and romanized data for twelve. An open GPU-accelerated pipeline for translation, quality assessment and repair, built with NeMo Data Designer, will be released alongside it.
Why it matters
Model builders get a substantial new resource for Indic pretraining. The 22 languages are the first languages of 97% of Indians, and the open pipeline lets developers generate more data from their own source documents.
- When
- 16 Oct 2026
- ID
WTA-N-2026-0306- Where
- India
- Who
- NVIDIA, IIT Madras, AI4Bharat, Bodhan AI; Mitesh Khapra
Summary by wheresthe.ai, from Hugging Face. Checked against the source on 16 Oct 2026.
Source copy: download the page as saved on 18 Oct 2026 (PDF). Document the source offers: document 1 (PDF). Also held by the Internet Archive.
Cite this story
WTA-N-2026-0306: “NVIDIA and IIT Madras announce 2.8 trillion-token Indic language corpus.” wheresthe.ai, 16 Oct 2026. https://wheresthe.ai/d/in-nvidia-and-iit-madras-announce-2-8-trillion-token-indic-l-2026-10/ (source: Hugging Face).
Spotted an error? Email [email protected].