WizardLM-2-7B: benchmark results
Microsoft's WizardLM-2 7B, a multilingual chat model built on Mistral-7B-v0.1, the smallest of the three WizardLM-2 family releases alongside the 70B and 8x22B models. Provider: Microsoft. Released 2024-04-16. Access: Open.
Aggregate Rating 1344 ± 19, rank #1395 of 1656 rated models, from 41 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open PL LLM - KLEJ NER (multiple choice, 5-shot) | 47.62 | Accuracy (%) | 55.8 |
| Open PL LLM - KLEJ NER (generative, 5-shot) | 46.55 | Accuracy (%) | 55.7 |
| Open PL LLM - PSC (multiple choice, 5-shot) | 67.16 | Binary F1 (%) | 54.2 |
| Open LLM Leaderboard - IFEval | 45.83 | Score | 51.4 |
| Open PL LLM - DYK (multiple choice, 5-shot) | 47.76 | Binary F1 (%) | 48.3 |
| Open PL LLM - Multiple Choice | 35.86 | Average Multiple-Choice Score (%) | 44.8 |
| Open PL LLM - PolEmo2-IN (generative, 5-shot) | 72.85 | Accuracy (%) | 44.2 |
| Open PL LLM - DYK (generative, 5-shot) | 47.53 | Binary F1 (%) | 43.7 |
| Open PL LLM - PolEmo2-OUT (multiple choice, 5-shot) | 66.4 | Accuracy (%) | 43.7 |
| Open PL LLM - 8TAGS (generative, 5-shot) | 69.76 | Accuracy (%) | 43.1 |
| Open PL LLM - PSC (generative, 5-shot) | 85.69 | Binary F1 (%) | 43.1 |
| Open PL LLM Leaderboard | 39.15 | Average Score (%) | 42.7 |
Interactive version: theaggregate.ai/model?slug=wizardlm-2-7b · How It Works · Data refreshed daily, snapshot 2026-10-20.