MATH-MC Level 5: leaderboard
MATH-MC Level 5 evaluates multiple-choice mathematical reasoning on the hardest MATH difficulty tier.
Metric: Accuracy (%). Source: github.com. Saturation forecast: Estimated already saturated. 69 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3.5 122B A10B | 99.92 | #173 |
| 2 | GPT-5.5 | 99.84 | #31 |
| 3 | Gemma 4 31B (IT) | 99.84 | #176 |
| 4 | Gemini 3.1 Pro (Preview) | 99.77 | #58 |
| 5 | Kimi K2.6 | 99.69 | #104 |
| 6 | Kimi K2 (Thinking) | 99.69 | #243 (Kimi K2) |
| 7 | GLM-4.6 FP8 | 99.69 | #278 |
| 8 | GLM-5.1 FP8 | 99.69 | #109 |
| 9 | Kimi K2.5 | 99.61 | #144 |
| 10 | Qwen 3.5 397B A17B | 99.61 | #146 |
| 11 | Qwen 3 30B A3B 2507 (Thinking) | 99.61 | #376 (Qwen 3 30B A3B 2507) |
| 12 | GPT-5 Mini | 99.53 | #186 |
| 13 | MiniMax-M2.5 | 99.53 | #303 |
| 14 | Step 3.5 Flash | 99.53 | #270 |
| 15 | DeepSeek V3.2 Speciale | 99.53 | #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-5 · How It Works · Data refreshed daily, snapshot 2026-10-20.