MATH-MC Level 3: leaderboard
MATH-MC Level 3 evaluates multiple-choice mathematical reasoning on intermediate MATH problems.
Metric: Accuracy (%). Source: github.com. Saturation forecast: Estimated already saturated. 69 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Kimi K2.5 | 99.73 | #144 |
| 2 | Gemini 3.1 Pro (Preview) | 99.64 | #58 |
| 3 | GPT-5.5 | 99.37 | #31 |
| 4 | Gemma 4 31B (IT) | 99.37 | #176 |
| 5 | Qwen 3.5 27B | 99.37 | #187 |
| 6 | Qwen 3.5 397B A17B | 99.28 | #146 |
| 7 | Qwen 3.6 Plus | 99.28 | #120 |
| 8 | Kimi K2 (Thinking) | 99.28 | #243 (Kimi K2) |
| 9 | Claude Opus 4.6 | 99.19 | #63 |
| 10 | Qwen 3.5 35B A3B | 99.19 | #256 |
| 11 | GLM-5 FP8 | 99.19 | #143 |
| 12 | NVIDIA-Nemotron-3-Super-120B-A12B-FP8 | 99.19 | #223 |
| 13 | Qwen 3.5 122B A10B | 99.1 | #173 |
| 14 | Gemma 4 26B A4B (IT) | 99.1 | #297 |
| 15 | DeepSeek V3.2 Speciale | 99.1 | #128 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=math-mc-level-3 · How It Works · Data refreshed daily, snapshot 2026-10-20.