OpenHathi 7B
India's first open Hindi LLM, built on Llama 2 by Sarvam AI.
About
OpenHathi-Hi-v0.1 is the first Hindi large language model in Sarvam AI's OpenHathi series, released in December 2023 — a week after the startup's $41M Series A. Built on Llama 2 7B with an extended 48,000-token bilingual tokenizer, it uses a two-phase training recipe: embedding alignment for Hindi followed by bilingual language modeling, developed with academic partners at AI4Bharat (IIT Madras).
Sarvam claimed GPT-3.5-level performance on Hindi tasks while retaining English ability, though these are vendor claims from a launch blog rather than independent benchmarks. As an early-2024 Hindi base model it is historically significant and useful for Indic-language research, but modern multilingual models have moved well past it on Hindi benchmarks.
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 72; Trending score 72/100 from researcher assessment.