LLM Releases
← Catalog

Sarvam-105B-1711

Available
Sarvam AIOpen source

Updated checkpoint of Sarvam-105B published on Hugging Face Nov 17 2026 under Apache 2.0 (NVFP4 build sarvamai/sarvam-105b-1711 and BF16 build sarvamai/sarvam-105b-1711-bf16). Same 105B-total / 10.3B-active MoE base as Sarvam-105B, further post-trained for tool calling, instruction following, agentic tasks and Indian languages. Sarvam's own runs report IFEval 96.5, tau2-Bench 85.5 and Indic Suite 79.1 (BF16, vendor-reported). Context window not stated on the card (evaluated at 131,072 tokens).

Specifications

License
Open source · Apache-2.0
Weights
Downloadable
Architecture
Mixture-of-Experts
Parameters
105B · 10.3B active
Context window
— tokens
Max output
—
Knowledge cutoff
—
Price (in / out, $/M)
—
Modalities
Text

Benchmarks

No benchmark scores recorded yet. Spotted some? Submit a correction.

Vendor-reported figures are claims until independently verified. See methodology.