Open PL LLM - PolQA Reranking (multiple choice, 5-shot): leaderboard
Choosing the passage relevant to a Polish PolQA question. Open PL LLM Leaderboard (SpeakLeash), 5-shot, multiple choice variant: the options are ranked by log-likelihood; accuracy (%), higher is better. Runs not available on the Hugging Face Hub are excluded.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Around February 2027. 319 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Bielik-11B-v2.1-Instruct | 85.63 | #679 |
| 2 | Bielik-11B-v2.2-Instruct | 85.4 | #669 |
| 3 | Bielik-11B-v2.3-Instruct | 84.66 | #700 |
| 4 | Qwen 2 72B | 84.6 | #515 |
| 5 | Qwen 2.5 32B Instruct | 84.37 | #509 |
| 6 | cogito-v1-preview-qwen-14B | 84.15 | |
| 7 | Qwen 2.5 32B | 83.53 | #544 |
| 8 | Qwen 2 72B Instruct | 83.15 | #579 |
| 9 | Llama 3.3 70B Instruct | 82.87 | #541 |
| 10 | Mistral Large 2 (Jul) | 82.78 | #490 |
| 11 | Qwen 3 14B | 82.57 | #542 |
| 12 | Bielik-11B-v2.0-Instruct | 82.45 | #750 |
| 13 | Llama 3.1 70B Instruct | 82.4 | #567 |
| 14 | Llama 3.1 Nemotron 70B Instruct | 82.37 | #563 |
| 15 | Qwen 2.5 14B Instruct | 82.19 | #654 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=open-pl-llm-polqa-reranking-multiple-choice-5-shot · How It Works · Data refreshed daily, snapshot 2026-11-05.