Open PL LLM - PolQA Closed Book (generative, 5-shot): leaderboard

Metric: Levenshtein Similarity (%). Source: huggingface.co. 319 models tracked.

Top models

#ModelScore
1Llama 3.1 405B Instruct FP885.98
2MSH-v1-Bielik-v2.3-Instruct-MedIT-merge85.77
3Mistral Large 2 (Jul)85.05
4Mistral Large 2 (Nov) Instruct (2411)85.05
5Bielik-11B-v285.05
6Bielik-11B-v2.3-Instruct84.11
7Bielik-11B-v2.2-Instruct83.59
8Llama 3.1 70B83.49
9Llama 4 Scout Base83.39
10Bielik-11B-v2.1-Instruct83.28
11Llama 3 70B83.07
12Bielik-11B-v2.5-Instruct82.97
13Bielik-11B-v2.0-Instruct82.45
14Mixtral-8x22B-v0.181.62
15Llama 3.3 70B Instruct80.37

Interactive version: theaggregate.ai/benchmark?slug=open-pl-llm-polqa-closed-book-generative-5-shot · How It Works · Data refreshed daily, snapshot 2026-09-19.