gemma-ko-7B-instruct-v0.62: benchmark results
lemon-mint's Gemma Ko 7B Instruct v0.62, a Korean-English chat fine-tune of openchat/openchat-3.5-0106-gemma (Gemma architecture). Provider: lemon-mint. Released 2024-03-01. Access: Open.
Aggregate Rating 1398 ± 19, rank #1208 of 1635 rated models, from 25 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Portuguese LLM - FaQuAD NLI | 80.26 | Macro F1 (%) | 89.7 |
| Open Portuguese LLM - ASSIN2 RTE | 93.13 | Macro F1 (%) | 86.6 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 42.99 | Exact match, flexible extract (%) | 78.4 |
| Open Korean LLM - Ko-EQ-Bench | 25.93 | EQ-Bench score | 72.3 |
| Open Korean LLM Leaderboard | 38.43 | Average Score (%) | 66.8 |
| Open Portuguese LLM - BLUEX | 55.49 | Accuracy (%) | 63.4 |
| Open Portuguese LLM - ENEM | 65.71 | Accuracy (%) | 62.5 |
| Open Portuguese LLM - OAB Exams | 43.51 | Accuracy (%) | 55.8 |
| Open Korean LLM - Ko-IFEval | 29.83 | Strict accuracy, prompt/instruction mean (%) | 53.1 |
| Open Medical LLM - MMLU Clinical Knowledge | 71.32 | Accuracy (%) | 51 |
| Open Medical LLM - MMLU Medical Genetics | 75 | Accuracy (%) | 50.3 |
| Open Medical LLM - PubMedQA | 73.8 | Accuracy (%) | 47.4 |
Interactive version: theaggregate.ai/model?slug=gemma-ko-7b-instruct-v0-62 · How It Works · Data refreshed daily, snapshot 2026-11-05.