Polish Knowledge Benchmark - Industry: leaderboard
Metric: Accuracy (%). Source: huggingface.co. 68 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Flash Lite (Preview) | 88.5 |
| 2 | Gemini 3.6 Flash | 88.5 |
| 3 | Gemini 3 Flash (Preview) | 87.86 |
| 4 | Gemini 3.1 Flash Lite | 87.86 |
| 5 | DeepSeek V4 Flash | 87.54 |
| 6 | Qwen 3.5 397B A17B | 87.22 |
| 7 | Gemini 2.5 Flash | 86.26 |
| 8 | Grok 4.1 Fast | 85.62 |
| 9 | Gemini 3.5 Flash Lite | 85.62 |
| 10 | DeepSeek V3 (0324) | 84.35 |
| 11 | DeepSeek V3.2 | 83.71 |
| 12 | Mistral Small 4 | 83.71 |
| 13 | Mistral Medium 3.1 | 83.39 |
| 14 | Qwen 3.5 122B A10B | 82.75 |
| 15 | Mistral Large 3 | 82.75 |
Interactive version: theaggregate.ai/benchmark?slug=polish-knowledge-benchmark-industry · How It Works · Data refreshed daily, snapshot 2026-09-19.