Open PL LLM - PoQuAD Open Book (generative, 5-shot): leaderboard
Answering PoQuAD reading-comprehension questions over a supplied Polish Wikipedia passage. Open PL LLM Leaderboard (SpeakLeash), 5-shot, generative variant: the answer is extracted from generated text; Levenshtein similarity to the reference answer (%), higher is better. Runs not available on the Hugging Face Hub are excluded.
Metric: Levenshtein Similarity (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 306 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Mistral Large 2 (Nov) Instruct (2411) | 76.25 | #480 |
| 2 | Mistral Small 3 | 73.99 | #630 |
| 3 | cogito-v1-preview-qwen-14B | 73.91 | |
| 4 | Qwen 2.5 72B | 73.79 | #512 |
| 5 | Mistral Large 2 (Jul) | 73.68 | #490 |
| 6 | Llama 3 70B | 73.66 | #632 |
| 7 | Llama 3.1 70B | 73.44 | #588 |
| 8 | Qwen 3 32B | 73.4 | #449 |
| 9 | QwQ 32B-Preview | 73.14 | #671 |
| 10 | Llama 3 70B Instruct | 72.94 | #640 |
| 11 | aya-expanse-32B | 72.74 | #687 |
| 12 | QwQ-32B | 72.48 | #422 |
| 13 | Qwen 2 72B | 72.43 | #515 |
| 14 | Qwen 2.5 72B Instruct | 72.41 | #469 |
| 15 | Llama 4 Scout Base | 72.38 | #708 |
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-poquad-open-book-generative-5-shot · How It Works · Data refreshed daily, snapshot 2026-11-05.