LLMZSZL Leaderboard — leaderboard
Polish-language LLM leaderboard comparing open and proprietary models on Polish zero-shot and language-understanding tasks.
Metric: Score. Source: huggingface.co. Status: saturation imminent. 99 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 2.5 72B Instruct | 69.06 |
| 2 | Qwen 2.5 72B | 68.5 |
| 3 | Mistral Large 2 (Jul) | 67.17 |
| 4 | Llama 3.3 70B Instruct | 67.13 |
| 5 | Qwen 3 32B | 66.84 |
| 6 | Mistral Large 2 (Nov) Instruct (2411) | 66.61 |
| 7 | Llama 3.1 70B Instruct | 66.59 |
| 8 | Llama 3 70B Instruct | 64.04 |
| 9 | Qwen 3 30B A3B | 63.52 |
| 10 | Llama 3.1 70B | 63.12 |
| 11 | Llama 3 70B | 62.22 |
| 12 | Qwen 2.5 32B | 61.04 |
| 13 | Qwen 3 14B | 61.02 |
| 14 | Mixtral 8x22B Instruct (v0.1) | 58.17 |
| 15 | Bielik-11B-v2.1-Instruct | 57.52 |
Interactive version: theaggregate.ai/benchmark?slug=llmzszl-leaderboard · How the rankings work · Data refreshed daily, snapshot 2026-07-22.