Benchmark
HMMT 2025
Math benchmarks measure how well models solve maths problems. Scores are percentages.
All math benchmarksScores from LLM Stats data.
- Models measured
- 32
- Scores recorded
- 32
- Sources
- 1
- IQ weight
- 2.39
Score distribution
- 82.9%
- Mean
- 90.7%
- Median
- 19.8%
- Std dev
- 38.8%
- 10th percentile
- 97.5%
- 90th percentile
- 28.9%
- Min
This benchmark has a weight of 2.39 for chat models, measured over 32 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.
Leaderboard
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.2 Pro· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 100% |
| 2 | GPT 5.2· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 99.4% |
| 3 | DeepSeek V3.2 Speciale· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 99.2% |
| 4 | Kimi K2 Thinking· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 97.5% |
| 5 | Qwen3.6 Plus· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 96.7% |
| 6 | Kimi K2.5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 95.4% |
| 7 | Qwen3.5 397B A17B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 94.8% |
| 8 | Nemotron 3 Super 120B A12B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 94.7% |
| 9 | GLM 5.2· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 94.4% |
| 10 | GLM 5.1· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 94% |
| 11 | Qwen3.6 27B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 93.8% |
| 12 | GPT 5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 93.3% |
| 13 | Grok 4 Fast· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 93.3% |
| 14 | Qwen3.5 27B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 92% |
| 15 | Qwen3.5 122B-A10B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 91.4% |
| 16 | Qwen3.6 35B A3B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 90.7% |
| 17 | DeepSeek Reasoner· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 90.2% |
| 18 | DeepSeek V3.2· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 90.2% |
| 19 | Qwen3.5 35B A3B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 89% |
| 20 | GPT-5 mini· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 87.8% |
| 21 | Sarvam 105B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 85.8% |
| 22 | MiMo V2 Flash· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 84.4% |
| 23 | DeepSeek V3.2 Exp· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 83.6% |
| 24 | Qwen3.5 9B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 83.2% |
| 25 | DeepSeek R1· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 79.4% |
| 26 | GPT-5 nano· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 75.6% |
| 27 | Qwen3.5 4B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 74% |
| 28 | Sarvam-30B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 73.3% |
| 29 | Kimi K2 Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 38.8% |
| 30 | GPT-4.1 mini· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 35% |
| 31 | DeepSeek V3.1· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 33.5% |
| 32 | GPT 4.1· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 28.9% |
Data credits: Data by LLM Stats