EuroEval Polish NLU - Polemo2: leaderboard
The "Polemo2" subtask of EuroEval Polish NLU. Polish-language EuroEval natural-language-understanding suite, separating NLU task performance from the broader all-task EuroEval aggregate.
Metric: Matthews correlation coefficient (%). Source: euroeval.com. Saturation forecast: Estimated already saturated. 234 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Llama 3.1 70B | 97.77 | #592 |
| 2 | Gemma 4 31B (IT) | 96.92 | #176 |
| 3 | GLM-4.7 Flash | 95.48 | #509 |
| 4 | Bielik-11B-v2.3-Instruct | 95.27 | #693 |
| 5 | Mistral Small 3.2 | 95.17 | #605 |
| 6 | Voxtral-Small-24B-2507 | 95.11 | #581 |
| 7 | Mistral Small 3.1 | 94.91 | #610 |
| 8 | Llama 3.1 8B | 94.7 | #1151 |
| 9 | gemma-3-27B-pt | 94.61 | #502 |
| 10 | Magistral Small 1.2 | 94.56 | #570 |
| 11 | Qwen 3.5 35B A3B Base | 93.43 | #310 |
| 12 | Llama 3.3 70B Instruct | 93.31 | #534 |
| 13 | Ministral-3-14B-Instruct-2512 | 93.21 | #615 |
| 14 | Qwen 3 8B (Non-reasoning) | 92.97 | #674 (Qwen 3 8B) |
| 15 | gemma-3-4B-pt Base | 92.84 | #1134 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=euroeval-polish-nlu-polemo2 · How It Works · Data refreshed daily, snapshot 2026-10-20.