BenCzechMark - Factual Knowledge: leaderboard
BenCzechMark Factual Knowledge 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: Around February 2027. 79 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | DeepSeek V3 (0324) | 88.28 | #349 |
| 2 | DeepSeek R1 0528 | 86.08 | #232 |
| 3 | Llama 3.1 405B Instruct | 77.11 | #470 |
| 4 | DeepSeek V4 Flash (0731) | 76.74 | #84 |
| 5 | Gemma 4 31B (IT) | 75.09 | #178 |
| 6 | Llama 4 Maverick Instruct | 73.08 | #467 |
| 7 | Mistral Large 2 (Nov) Instruct (2411) | 72.34 | #480 |
| 8 | MiniMax M1 80k | 71.43 | #437 |
| 9 | c4ai-command-a-03-2025 | 69.6 | #468 |
| 10 | Qwen 2.5 72B | 69.23 | #512 |
| 11 | Llama 3.1 70B Instruct | 66.48 | #567 |
| 12 | Qwen 2.5 72B Instruct | 66.3 | #469 |
| 13 | Llama 3.3 70B Instruct | 65.93 | #541 |
| 14 | Muse Glimmer 30B | 65.93 | #191 |
| 15 | Llama 3.1 70B | 64.65 | #588 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=benczechmark-factual-knowledge · How It Works · Data refreshed daily, snapshot 2026-11-05.