- Math500 98.6
- Live Code Bench v6 71.7
- MMLU 90.6
- MMLU Pro 81.7
- Writing Bench (Hugging Face card table, official critic model) 80.5
- Arena Hard v2 71
- IF Eval 84.8
- GPQA Diamond 78.7
- AIME 25 (88.3 without tools, 96.7 with tools) 88.3
- HMMT (Feb 25) 85.8
- HMMT (Nov 25) 85.8
- Beyond AIME 69.1
- BrowseComp 49.5
- SWE Bench Verified (SWE-Agent Harness) 45
- Tau2 (avg.) 68.3
- HLE 11.2
- Average pairwise win rate across all dimensions on the Indic benchmark (indivibe) 90
Sarvam 105B
Sarvam 105B places in the top three on MATH-500.
16
Benchmarks with a score
0
Ranked first, out of five or more models
1
In the top three
Sarvam AI Mar 2026
Sarvam 105B
Open weights Flagship
139.0
#185 of 404
Self-Reported Index, from the lab's own numbers. No Epoch Capabilities Index score yet.
- BrowseComp 49.5% #63 /81
- tau2-bench 68.3% #60 /79
- MMLU-Pro 81.7% #64 /167
Where it ranks
Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.
Knowledge and language
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| MMLU | 90.6% | #9 of 131 | Lab model card | |
| MMLU-Pro | 81.7% | #64 of 167 | Lab model card | |
| Arena-Hard v2.0 | 71% | #11 of 24 | Lab model card | |
| IndiVibe | 90% | #1 of 2 | Lab model card | |
| IFEval | 84.8% | #84 of 135 | Lab model card | |
| Humanity's Last Exam | 11.2% | #134 of 154 | Lab model card |
Math and science
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| MATH-500 | 98.6% | #2 of 70 | Lab model card | |
| AIME 2025 | 88.3% | #59 of 143 | Lab model card | |
| GPQA Diamond | 78.7% | #127 of 302 | Lab model card | |
| Beyond AIME | 69.1% | #6 of 14 | Lab model card | |
| HMMT 2025 | 85.8% | #30 of 54 | Lab model card | |
| HMMT Nov 2025 | 85.8% | #20 of 28 | Lab model card |
Coding
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| LiveCodeBench v6 | 71.7% | #53 of 90 | Lab model card | |
| SWE-bench Verified | 45% | #137 of 155 | Lab model card |
Agents and tools
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| tau2-bench | 68.3% | #60 of 79 | Lab model card | |
| BrowseComp | 49.5% | #63 of 81 | Lab model card |
Its lab's numbers put it at 139.0, number 185 of 404
The Self-Reported Index fits one score to the 15 capability results Sarvam AI reported for this model, on the scale of the Epoch Capabilities Index. Labs choose which benchmarks to show and how to run them, so read it as the lab's claim, not an independent measurement. The 90% interval runs from 137.0 to 141.0. How we compute it.
- MATH-500 98.6
- MMLU 90.6
- AIME 2025 88.3
- HMMT 2025 85.8
- HMMT Nov 2025 85.8
- IFEval 84.8
- MMLU-Pro 81.7
- GPQA Diamond 78.7
- LiveCodeBench v6 71.7
- Arena-Hard v2.0 71
- Beyond AIME 69.1
- tau2-bench 68.3
- BrowseComp 49.5
- SWE-bench Verified 45
- Humanity's Last Exam 11.2
Its launch post
The benchmarks the lab chose to headline, with the scores it gave.