Polish Knowledge Benchmark - Overall: leaderboard
Polish Knowledge Benchmark (PolishQA): 15,000 synthetic four-option questions about Poland generated by DeepSeek 3.2 and difficulty-rated by Qwen3.5-397B-A17B, built from SpeakLeash community sources with context-dependent questions filtered out. Accuracy is correct over answered-question totals across all difficulty levels, 0-100, higher is better.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 69 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Flash (Preview) | 92.62 | #90 |
| 2 | Gemini 3.6 Flash | 91.38 | #57 |
| 3 | Gemini 3.1 Flash Lite (Preview) | 90.07 | #201 |
| 4 | Gemini 3.1 Flash Lite | 89.92 | #200 |
| 5 | DeepSeek V4 Flash | 88.91 | #155 |
| 6 | Gemini 3.5 Flash Lite | 88.37 | #174 |
| 7 | Qwen 3.5 397B A17B | 87.16 | #154 |
| 8 | Gemini 2.5 Flash | 87.15 | #255 |
| 9 | Mistral Large 3 | 87.09 | #410 |
| 10 | Mistral Large 4 | 86.79 | #106 |
| 11 | Grok 4.1 Fast | 86.33 | #227 |
| 12 | DeepSeek V3.2 | 86.06 | #210 |
| 13 | Mistral Medium 3.1 | 85.69 | #418 |
| 14 | Gemini 2.0 Flash (001) | 85.41 | #365 |
| 15 | Gemma 4 31B (IT) | 84.73 | #178 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=polish-knowledge-benchmark-overall · How It Works · Data refreshed daily, snapshot 2026-11-05.