EuroEval Czech NLU - PONER: leaderboard
The "PONER" subtask of EuroEval Czech NLU. Czech-language EuroEval natural-language-understanding suite, separating NLU task performance from the broader all-task EuroEval aggregate.
Metric: Micro F1 excluding MISC (%). Source: euroeval.com. Saturation forecast: Around January 2027. 229 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Sonnet 4.5 (Thinking) | 69.75 | #142 (Claude Sonnet 4.5) |
| 2 | Qwen 3 235B A22B 2507 FP8 (Thinking) | 66.66 | #246 (Qwen 3 235B A22B 2507 FP8) |
| 3 | GPT-5.4 Mini (2026-03-17) (High) | 65.91 | #169 (GPT-5.4 Mini (2026-03-17)) |
| 4 | GPT-5 | 65.66 | #98 |
| 5 | Grok 4.20 Beta (0309) (Reasoning) | 63.6 | #204 (Grok 4.20 Beta (0309)) |
| 6 | Gemini 3 Flash (Preview) (Thinking) | 63.36 | #82 (Gemini 3 Flash (Preview)) |
| 7 | Mistral Small 3.1 | 62.66 | #610 |
| 8 | Qwen 3 4B 2507 (Thinking) | 62.54 | #538 (Qwen 3 4B 2507) |
| 9 | Mistral Small 3.2 | 62.28 | #605 |
| 10 | Voxtral-Small-24B-2507 | 61.89 | #581 |
| 11 | Qwen 3.5 27B FP8 | 60.46 | #164 |
| 12 | Magistral Small 1.2 | 60.05 | #570 |
| 13 | gemma-3-27B-pt | 59.59 | #502 |
| 14 | NVIDIA-Nemotron-3-Nano-30B-A3B-BF16 | 59.57 | #679 |
| 15 | Llama 3.1 70B | 59.01 | #592 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=euroeval-czech-nlu-poner · How It Works · Data refreshed daily, snapshot 2026-10-20.