Sarvam AIopen-weight
Released: 2025-05-20Verified

Sarvam-M

24B open-weights hybrid reasoning model from Sarvam AI, post-trained from Mistral-Small-3.1-24B-Base-2503 on English and eleven major Indic languages. Introduced in May 2025 as Sarvam's first public chat model; Sarvam's own write-up notes it trails the base model by about one point on English knowledge benchmarks such as MMLU.

Intelligence Index
—/ 100
Artificial Analysis Intelligence Index (independent composite)
Throughput Speed
—tok/sec
Output streaming throughput
Time to First Answer
—ms
Time to first answer token, including any thinking
LiveBench Contam-Free
—%
Monthly suite score

Sourced Benchmark Evaluations (0)

Verified performance across authoritative benchmarks with provenance tracking.
No verified benchmark scores recorded for this model yet.
Local Hardware Execution (Ollama)Hardware Compatible
# 1. Pull and execute model locally with Ollama
ollama run sarvam-m

# 2. Or invoke via local OpenAI-compatible endpoint
curl http://localhost:11434/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "sarvam-m",
    "messages": [{"role": "user", "content": "Analyze reasoning chains on AIME 2026"}]
  }'

Evaluation & Sourcing Notes

Scores listed for Sarvam-M represent verified evaluations extracted from official research papers, independent evaluation suites (HELM, LMSYS, OpenCompass, LiveBench), and verified audit reports.

All benchmarks marked as saturated or deprecated reflect historical performance where the benchmark no longer provides active discriminative power.