CPTU Bench — leaderboard
Polish language comprehension benchmark: evaluates LLMs on sentiment analysis, language understanding, phraseology, and tricky questions in Polish.
Metric: Average Score (1-5). Source: huggingface.co. Status: saturation imminent. 93 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 2.0 Flash (001) | 4.29 |
| 2 | DeepSeek V3.2 | 4.14 |
| 3 | DeepSeek R1 | 4.14 |
| 4 | Gemini 2.0 Flash Lite (001) | 4.09 |
| 5 | DeepSeek V3.1 | 4.03 |
| 6 | DeepSeek V3 (0324) | 4.03 |
| 7 | DeepSeek V3 | 4.02 |
| 8 | Mistral Large 2 (Nov) Instruct (2411) | 4 |
| 9 | Kimi K2 0905 | 3.98 |
| 10 | Qwen 2.5 72B Instruct | 3.95 |
| 11 | Mistral Large 2 (Jul) | 3.93 |
| 12 | Llama 4 Maverick Instruct FP8 | 3.93 |
| 13 | Mistral Small 3.1 | 3.9 |
| 14 | Mistral Small 3.2 | 3.83 |
| 15 | GPT-OSS-120B | 3.82 |
Interactive version: theaggregate.ai/benchmark?slug=cptu-bench · How the rankings work · Data refreshed daily, snapshot 2026-07-22.