Llama 3 8B Instruct V0.10: benchmark results
MaziyarPanahi's iterative general-chat fine-tune of Meta Llama 3 8B Instruct, built on top of his v0.9 model. Provider: MaziyarPanahi. Released 2024-06-04. Access: Open.
Aggregate Rating 1475 ± 17, rank #868 of 1656 rated models, from 29 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Korean LLM - KorNAT-Helpful | 54.05 | Normalized accuracy (%) | 94.6 |
| Open Korean LLM - Ko-IFEval | 49.9 | Strict accuracy, prompt/instruction mean (%) | 93.4 |
| Open LLM Leaderboard - IFEval | 76.67 | Score | 92.4 |
| Open Korean LLM Leaderboard | 44.61 | Average Score (%) | 90.6 |
| Open Ko-LLM Leaderboard | 44.61 | Average (%) | 90.4 |
| Open Korean LLM - Ko-EQ-Bench | 37.05 | EQ-Bench score | 87.8 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 51.4 | Exact match, flexible extract (%) | 85 |
| Open Korean LLM - KorNAT-Harmless | 65.99 | Normalized accuracy (%) | 83.7 |
| Open Korean LLM - Ko-Winogrande | 65.04 | Accuracy (%) | 82.2 |
| Open Portuguese LLM - ENEM | 71.38 | Accuracy (%) | 80.2 |
| Open Chinese LLM - TruthfulQA MC | 58.61 | Accuracy (%) | 77.5 |
| Open Portuguese LLM - OAB Exams | 50.3 | Accuracy (%) | 76.6 |
Interactive version: theaggregate.ai/model?slug=llama-3-8b-instruct-v0-10 · How It Works · Data refreshed daily, snapshot 2026-10-20.