Benchmark
U-MATH integral_calc accuracy
Math benchmarks measure how well models solve maths problems. Scores are percentages. Higher is better.
- Models measured
- 30
- Scores recorded
- 90
- Sources
- 1
- IQ weight
- 2.00
Score distribution
- 32.3%
- Mean
- 27.3%
- Median
- 22.9%
- Std dev
- 9.6%
- 10th percentile
- 67.3%
- 90th percentile
- 1.3%
- Min
This benchmark has a weight of 2.00 for chat models, measured over 30 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 | o1integral_calc, text, gpt-4o-2024-08-06· U-MATH and mu-MATH | 90.7% |
| 2 | o3 Miniintegral_calc, text, gpt-4o-2024-08-06· U-MATH and mu-MATH | 90.7% |
| 3 | Gemini 2.0 Flash Experimentalintegral_calc, text, gpt-4o-2024-08-06, thinking· U-MATH and mu-MATH | 88.7% |
| 4 | DeepSeek R1integral_calc, text, gpt-4o-2024-08-06· U-MATH and mu-MATH | 87.3% |
| 5 | o1integral_calc, all, gpt-4o-2024-08-06· U-MATH and mu-MATH | 83.2% |
| 6 | Gemini 2.0 Flash Experimentalintegral_calc, all, gpt-4o-2024-08-06, thinking· U-MATH and mu-MATH | 82.2% |
| 7 | o3 Miniintegral_calc, all, gpt-4o-2024-08-06· U-MATH and mu-MATH | 82.2% |
| 8 | DeepSeek R1integral_calc, all, gpt-4o-2024-08-06· U-MATH and mu-MATH | 76.9% |
| 9 | O1 Miniintegral_calc, text, gpt-4o-2024-08-06· U-MATH and mu-MATH | 72% |
| 10 | QwQ 32B Previewintegral_calc, text, gpt-4o-2024-08-06· U-MATH and mu-MATH | 67.3% |
| 11 | Gemini 2.0 Flash Experimentalintegral_calc, visual, gpt-4o-2024-08-06, thinking· U-MATH and mu-MATH | 65.5% |
| 12 | O1 Miniintegral_calc, all, gpt-4o-2024-08-06· U-MATH and mu-MATH | 64.9% |
| 13 | o1integral_calc, visual, gpt-4o-2024-08-06· U-MATH and mu-MATH | 63.8% |
| 14 | QwQ 32B Previewintegral_calc, all, gpt-4o-2024-08-06· U-MATH and mu-MATH | 62.5% |
| 15 | o3 Miniintegral_calc, visual, gpt-4o-2024-08-06· U-MATH and mu-MATH | 60.3% |
| 16 | QVQ 72B Previewintegral_calc, visual, gpt-4o-2024-08-06· U-MATH and mu-MATH | 55.2% |
| 17 | DeepSeek R1integral_calc, visual, gpt-4o-2024-08-06· U-MATH and mu-MATH | 50% |
| 18 | QwQ 32B Previewintegral_calc, visual, gpt-4o-2024-08-06· U-MATH and mu-MATH | 50% |
| 19 | O1 Miniintegral_calc, visual, gpt-4o-2024-08-06· U-MATH and mu-MATH | 46.6% |
| 20 | QVQ 72B Previewintegral_calc, all, gpt-4o-2024-08-06· U-MATH and mu-MATH | 45.2% |
| 21 | Qwen2.5 Math 72B Instructintegral_calc, text, gpt-4o-2024-08-06· U-MATH and mu-MATH | 44% |
| 22 | QVQ 72B Previewintegral_calc, text, gpt-4o-2024-08-06· U-MATH and mu-MATH | 41.3% |
| 23 | DeepSeek V3integral_calc, visual, gpt-4o-2024-08-06· U-MATH and mu-MATH | 39.7% |
| 24 | DeepSeek V3integral_calc, all, gpt-4o-2024-08-06· U-MATH and mu-MATH | 38.9% |
| 25 | Qwen2.5 Math 72B Instructintegral_calc, all, gpt-4o-2024-08-06· U-MATH and mu-MATH | 38.9% |
| 26 | DeepSeek V3integral_calc, text, gpt-4o-2024-08-06· U-MATH and mu-MATH | 38.7% |
| 27 | Mistral Large Instruct 2411integral_calc, text, gpt-4o-2024-08-06· U-MATH and mu-MATH | 36.7% |
| 28 | Phi 4integral_calc, visual, gpt-4o-2024-08-06· U-MATH and mu-MATH | 36.2% |
| 29 | Qwen2.5 72B Instructintegral_calc, text, gpt-4o-2024-08-06· U-MATH and mu-MATH | 35.3% |
| 30 | Qwen2.5 32B Instructintegral_calc, text, gpt-4o-2024-08-06· U-MATH and mu-MATH | 34.7% |
| 31 | Qwen2.5 32B Instructintegral_calc, all, gpt-4o-2024-08-06· U-MATH and mu-MATH | 32.2% |
| 32 | Qwen2.5 7B Instructintegral_calc, text, gpt-4o-2024-08-06· U-MATH and mu-MATH | 32% |
| 33 | Gemini 1.5 Flashintegral_calc, visual, gpt-4o-2024-08-06· U-MATH and mu-MATH | 31% |
| 34 | Gemini 1.5 Flashintegral_calc, all, gpt-4o-2024-08-06· U-MATH and mu-MATH | 30.8% |
| 35 | Mistral Large Instruct 2411integral_calc, all, gpt-4o-2024-08-06· U-MATH and mu-MATH | 30.8% |
| 36 | Gemini 1.5 Flashintegral_calc, text, gpt-4o-2024-08-06· U-MATH and mu-MATH | 30.7% |
| 37 | Qwen2.5 72B Instructintegral_calc, all, gpt-4o-2024-08-06· U-MATH and mu-MATH | 30.3% |
| 38 | Phi 4integral_calc, all, gpt-4o-2024-08-06· U-MATH and mu-MATH | 28.4% |
| 39 | Qwen2.5 7B Instructintegral_calc, all, gpt-4o-2024-08-06· U-MATH and mu-MATH | 28.4% |
| 40 | GPT 4ointegral_calc, visual, gpt-4o-2024-08-06· U-MATH and mu-MATH | 27.6% |
| 41 | Qwen2-VL (72B) Instructintegral_calc, visual, gpt-4o-2024-08-06· U-MATH and mu-MATH | 27.6% |
| 42 | GPT 4ointegral_calc, all, gpt-4o-2024-08-06· U-MATH and mu-MATH | 27.4% |
| 43 | GPT 4ointegral_calc, text, gpt-4o-2024-08-06· U-MATH and mu-MATH | 27.3% |
| 44 | Llama 3.3 70B Instructintegral_calc, text, gpt-4o-2024-08-06· U-MATH and mu-MATH | 27.3% |
| 45 | Gemini 1.5 Prointegral_calc, text, gpt-4o-2024-08-06· U-MATH and mu-MATH | 27.3% |
| 46 | Gemini 1.5 Prointegral_calc, all, gpt-4o-2024-08-06· U-MATH and mu-MATH | 26.4% |
| 47 | Qwen2.5 32B Instructintegral_calc, visual, gpt-4o-2024-08-06· U-MATH and mu-MATH | 25.9% |
| 48 | Qwen2.5 Math 72B Instructintegral_calc, visual, gpt-4o-2024-08-06· U-MATH and mu-MATH | 25.9% |
| 49 | Llama 3.3 70B Instructintegral_calc, all, gpt-4o-2024-08-06· U-MATH and mu-MATH | 25.5% |
| 50 | Phi 4integral_calc, text, gpt-4o-2024-08-06· U-MATH and mu-MATH | 25.3% |
Showing 1–50 of 90 results.
Data credits: Toloka, U-MATH and mu-MATH public leaderboard. MIT Space; generated solution outputs retain underlying model terms and are not ingested. (licence)