EuroEval Czech NLU - CSFD Sentiment: leaderboard
The "CSFD Sentiment" subtask of EuroEval Czech NLU. Czech-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. 229 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | gemma-3-12B-pt | 75.96 | #686 |
| 2 | gemma-3-27B-pt | 72.91 | #502 |
| 3 | gemma-3-4B-pt Base | 68.79 | #1134 |
| 4 | lapa-v0.1.2-instruct | 66.23 | #672 |
| 5 | Llama 3.1 70B | 66.16 | #592 |
| 6 | EuroLLM-9B | 64.86 | #1254 |
| 7 | Llama 3.1 8B | 64.67 | #1151 |
| 8 | Gemma 3 12B (IT) | 62.6 | #664 |
| 9 | MamayLM-Gemma-3-12B-IT-v1.0 | 62.56 | #584 |
| 10 | Qwen 2.5 72B Instruct | 61.48 | #454 |
| 11 | Llama 3.3 70B Instruct | 61.31 | #534 |
| 12 | Magistral Small 1.2 | 60.91 | #570 |
| 13 | EuroLLM-9B-Instruct | 60.62 | #1054 |
| 14 | Voxtral-Small-24B-2507 | 60.61 | #581 |
| 15 | Ministral-3-14B-Instruct-2512 | 60.48 | #615 |
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-csfd-sentiment · How It Works · Data refreshed daily, snapshot 2026-10-20.