zephyr-all-7B: benchmark results
Provider: HuggingFace. Access: Open.
Unified ELO 1342 ± 17, rank #2275 of 2656 rated models, from 10 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Korean LLM - Ko-GPQA-Diamond | 24.24 | Normalized accuracy (%) | 62.4 |
| Open Korean LLM - Ko-EQ-Bench | 0.46 | EQ-Bench score | 45 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 6.07 | Exact match, flexible extract (%) | 26.6 |
| Open Korean LLM - Ko-IFEval | 23.03 | Strict accuracy, prompt/instruction mean (%) | 25.6 |
| Open Korean LLM Leaderboard | 30.07 | Average Score (%) | 16.9 |
| Open Ko-LLM Leaderboard | 30.07 | Average (%) | 16.8 |
| Open Korean LLM - KorNAT-CKA | 17.74 | Normalized accuracy (%) | 16.4 |
| Open Korean LLM - Ko-Winogrande | 53.51 | Accuracy (%) | 12.9 |
| Open Korean LLM - KorNAT-Helpful | 41.25 | Normalized accuracy (%) | 11.1 |
| Open Korean LLM - KorNAT-Harmless | 59.45 | Normalized accuracy (%) | 6.6 |
Interactive version: theaggregate.ai/model?slug=zephyr-all-7b · How It Works · Data refreshed daily, snapshot 2026-09-19.