Bielik-11B-v2.2-Instruct: benchmark results
SpeakLeash's Polish-focused 11B instruct fine-tune of the Bielik v2 base, a Mistral 7B v0.2 depth-upscaled to 11B (August 2024). Provider: SpeakLeash. Released 2024-08-26. Access: Open.
Aggregate Rating 1535 ± 15, rank #660 of 1656 rated models, from 82 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open PL LLM - RAG | 76.92 | Average RAG Score (%) | 100 |
| Open PL LLM - PolQA Reranking (multiple choice, 5-shot) | 85.4 | Accuracy (%) | 98.4 |
| Open PL LLM - DYK (generative, 5-shot) | 71.53 | Binary F1 (%) | 97.8 |
| Open PL LLM - PolQA Open Book (generative, 5-shot) | 92.12 | Levenshtein Similarity (%) | 97.5 |
| Open PL LLM - Multiple Choice | 63.41 | Average Multiple-Choice Score (%) | 96.3 |
| Open PL LLM - PolQA Closed Book (generative, 5-shot) | 83.59 | Levenshtein Similarity (%) | 95 |
| Open PL LLM - PolEmo2-IN (multiple choice, 5-shot) | 87.26 | Accuracy (%) | 94.7 |
| Open PL LLM - DYK (multiple choice, 5-shot) | 70.24 | Binary F1 (%) | 93.2 |
| Open PL LLM - PSC (multiple choice, 5-shot) | 91.68 | Binary F1 (%) | 93.1 |
| Open PL LLM Leaderboard | 65.57 | Average Score (%) | 93 |
| Open PL LLM - PolEmo2-IN (generative, 5-shot) | 86.43 | Accuracy (%) | 92.2 |
| Open Portuguese LLM - ASSIN2 RTE | 93.67 | Macro F1 (%) | 91.2 |
Interactive version: theaggregate.ai/model?slug=bielik-11b-v2-2-instruct · How It Works · Data refreshed daily, snapshot 2026-10-20.