MATH-MC Level 1: leaderboard
MATH-MC Level 1 evaluates multiple-choice mathematical reasoning on the easiest MATH difficulty tier.
Metric: Accuracy (%). Source: github.com. Saturation forecast: Estimated already saturated. 69 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Kimi K2.5 | 99.3 | #144 |
| 2 | Kimi K2.6 | 99.3 | #104 |
| 3 | MiMo-V2-Flash | 99.3 | #372 |
| 4 | NVIDIA-Nemotron-3-Super-120B-A12B-FP8 | 99.3 | #223 |
| 5 | Gemini 3.1 Pro (Preview) | 99.07 | #58 |
| 6 | Gemma 4 31B (IT) | 99.07 | #176 |
| 7 | Step 3.5 Flash | 99.07 | #270 |
| 8 | GPT-OSS-120B | 98.84 | #341 |
| 9 | Qwen 3 Next 80B A3B (Thinking) | 98.84 | #320 (Qwen 3 Next 80B A3B) |
| 10 | MiniMax-M2.5 | 98.84 | #303 |
| 11 | OLMo 3.1 32B (Thinking) | 98.84 | #632 (OLMo 3.1 32B) |
| 12 | DeepSeek V3.2 Speciale | 98.84 | #128 |
| 13 | GLM-4.7 FP8 | 98.84 | #209 |
| 14 | GLM-5.1 FP8 | 98.84 | #109 |
| 15 | GPT-5.5 | 98.6 | #31 |
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-1 · How It Works · Data refreshed daily, snapshot 2026-10-20.