Llama 3 8B Instruct V0.8: benchmark results
MaziyarPanahi's Axolotl fine-tune of Meta Llama 3 8B Instruct, an iterative refinement built on his earlier v0.4 model. Provider: Meta. Released 2024-05-01. Access: Open.
Unified ELO 1530 ± 1, rank #547 of 1392 rated models, from 15 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard - IFEval | 75.28 | Score | 90.6 |
| Open Korean LLM Leaderboard | 42.97 | Average Score (%) | 85.2 |
| Open Chinese LLM - TruthfulQA MC | 57.72 | Accuracy (%) | 73.9 |
| Open LLM Leaderboard - MMLU-Pro | 31.7 | Score | 68.3 |
| Open LLM Leaderboard - GPQA | 7.38 | Score | 62.2 |
| Open Chinese LLM - ARC Challenge | 54.01 | Accuracy (%) | 61.9 |
| Open Chinese LLM Leaderboard | 57.92 | Average Score (%) | 61.4 |
| Open Chinese LLM - GSM8K | 47.76 | Accuracy (%) | 59.6 |
| Open Chinese LLM - CMMLU | 54.32 | Accuracy (%) | 56.1 |
| Open LLM Leaderboard - MuSR | 10.92 | Score | 55.9 |
| Open Chinese LLM - C-Eval Semantic | 70.85 | Accuracy (%) | 54 |
| Open LLM Leaderboard - BBH | 28.27 | Score | 46.1 |
Interactive version: theaggregate.ai/model?slug=llama-3-8b-instruct-v0-8 · How It Works · Data refreshed daily, snapshot 2026-09-05.