recoilme-gemma-2-9B-v0.2: benchmark results
recoilme's mergekit blend of six Gemma 2 9B fine-tunes, an earlier v0.2 iteration of the recoilme-gemma-2 merge series. Provider: Other. Released 2024-09-18. Access: Open.
Aggregate Rating 1544 ± 18, rank #629 of 1656 rated models, from 144 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Korean LLM - Ko-IFEval | 69.49 | Strict accuracy, prompt/instruction mean (%) | 99.7 |
| Open Ko-LLM Leaderboard | 53.44 | Average (%) | 99 |
| Open Korean LLM Leaderboard | 53.44 | Average Score (%) | 99 |
| Open Korean LLM - Ko-EQ-Bench | 58.92 | EQ-Bench score | 98.3 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 69.37 | Exact match, flexible extract (%) | 98 |
| Open LLM Leaderboard - IFEval | 75.92 | Score | 91.3 |
| Open Arabic LLM - Aratrust Unfairness | 94.55 | Accuracy (%) | 91 |
| Open Portuguese LLM - ASSIN2 RTE | 93.66 | Macro F1 (%) | 91 |
| Open LLM Leaderboard - MuSR | 46.86 | Score | 90.8 |
| Open Korean LLM - KorNAT-CKA | 32.84 | Normalized accuracy (%) | 89.8 |
| Open Portuguese LLM - ENEM | 75.23 | Accuracy (%) | 88.9 |
| Open Arabic LLM - Aratrust Offensive | 92.75 | Accuracy (%) | 88.2 |
Interactive version: theaggregate.ai/model?slug=recoilme-gemma-2-9b-v0-2 · How It Works · Data refreshed daily, snapshot 2026-10-20.