Sarvam 30B
SarvamOpen weightsReasoningProvisionalReleased Mar 6, 2026
Overall score
28.6
Rank 190 of 222 · 14.3–43.0
Context window
64K
64,000 tokens
Price per 1M tokens
Free in · Free out
Source: BenchLM
Blended price
Free
3:1 input:output tokens
Output speed
—
Not measured yet
Category scores
Scores per capability area with the model's rank among ranked models in that category.
- 1Knowledge64.9
- 2Instruction following#14028.8
- 3Reasoning25.0
- 4Coding24.5
- 5Agentic & tool use18.4
Benchmark results (8)
As reported in system cards, papers and public evaluations. Position compares this model with every other model reporting the same benchmark.
Math
| MATH-500 | 97.0 | #6 of 20 |
| HMMT Feb 2025 | 73.3 | #13 of 13 |
| HMMT Nov 2025 | 74.2 | #12 of 12 |
Coding
| SWE-bench Verified | 34.0 | #95 of 98 |
| HumanEval | 92.1 | #2 of 18 |
Knowledge
| MMLU-Pro | 80.0 | #35 of 67 |
| MMLU | 85.1 | #16 of 28 |
Agentic & tool use
| BrowseComp | 35.5 | #50 of 50 |
Pricing by source
USD per 1M tokens.
| Source | Input | Output | Cached in |
|---|---|---|---|
| List price (BenchLM) | Free | Free | — |
Sarvam's official API pricing page lists Sarvam 30B as free per token.
Speed
Measured by us through OpenRouter (median of trials).
Not available on OpenRouter, so we cannot measure it yet.
Related models
Same family
- Sarvam 105B31.7
Similar score
- Gemma 3 27B28.6 · $0.15
- Qwen2.5-72B28.7 · —
- Mistral Medium 328.8 · $0.80
- Granite-4.0-H-1B28.8 · —
- Granite-4.0-350M28.5 · —