Polish Knowledge Benchmark - Transport & Communication: leaderboard
Topic category of the Polish Knowledge Benchmark (PolishQA) covering transport and communication in Poland. The benchmark has 15,000 synthetic four-option questions about Poland generated by DeepSeek 3.2 and difficulty-rated by Qwen3.5-397B-A17B. Accuracy is correct over total in the category, 0-100, higher is better.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Around February 2027. 69 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Flash (Preview) | 86.73 | #90 |
| 2 | Gemini 3.6 Flash | 86 | #57 |
| 3 | DeepSeek V4 Flash | 84.28 | #155 |
| 4 | Gemini 3.1 Flash Lite | 84.28 | #200 |
| 5 | Gemini 3.1 Flash Lite (Preview) | 84.03 | #201 |
| 6 | Qwen 3.5 397B A17B | 83.05 | #154 |
| 7 | Gemini 3.5 Flash Lite | 80.34 | #174 |
| 8 | Gemini 2.5 Flash | 80.1 | #255 |
| 9 | Grok 4.1 Fast | 80.1 | #227 |
| 10 | Mistral Large 3 | 79.85 | #410 |
| 11 | Mistral Large 4 | 79.12 | #106 |
| 12 | Gemma 4 31B (IT) | 78.38 | #178 |
| 13 | DeepSeek V3.2 | 77.4 | #210 |
| 14 | Gemini 2.0 Flash (001) | 77.4 | #365 |
| 15 | Gemma 4 26B A4B (IT) | 76.9 | #293 |
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-transport-communication · How It Works · Data refreshed daily, snapshot 2026-11-05.