BenCzechMark - Czech Math Reasoning: leaderboard
BenCzechMark Czech Math Reasoning category of Czech tasks. Scores are duel win scores (0-100): the share of pairwise statistical duels a model wins against every other model on the category's tasks, averaged over tasks, so they shift as the roster changes; higher is better.
Metric: Duel Win Score (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 79 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | DeepSeek V4 Flash (0731) | 93.59 | #84 |
| 2 | DeepSeek V3 (0324) | 92.63 | #349 |
| 3 | Gemma 4 31B (IT) | 90.71 | #178 |
| 4 | DeepSeek R1 0528 | 87.5 | #232 |
| 5 | Llama 4 Maverick Instruct | 86.86 | #467 |
| 6 | Qwen 2.5 72B | 84.62 | #512 |
| 7 | Qwen 2.5 72B Instruct | 82.37 | #469 |
| 8 | Qwen 2.5 32B Instruct | 79.17 | #509 |
| 9 | Llama 3.1 405B Instruct | 76.6 | #470 |
| 10 | Gemma 4 26B A4B (IT) | 72.12 | #293 |
| 11 | Qwen 3.6 27B | 71.15 | #182 |
| 12 | MiniMax M1 80k | 69.55 | #437 |
| 13 | c4ai-command-a-03-2025 | 67.63 | #468 |
| 14 | Qwen 2 72B Instruct | 66.35 | #579 |
| 15 | Mistral Large 2 (Nov) Instruct (2411) | 64.1 | #480 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=benczechmark-czech-math-reasoning · How It Works · Data refreshed daily, snapshot 2026-11-05.