Llama 3 70B: benchmark results
Meta's 70B pretrained base model from the original Llama 3 release, with an 8K context - the foundation for its instruct variant (April 2024). Provider: Meta. Released 2024-04-18. Access: Open.
Unified ELO 1535 ± 1, rank #512 of 1392 rated models, from 186 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| CyberSecEval-3 | 44.14 | Score (%) | 100 |
| IFEval Leaderboard | 83.31 | Final Score | 100 |
| Open Japanese LLM - MR | 94.8 | Score (%) | 99.2 |
| HELM NarrativeQA | 79.82 | F1 (%) | 98.9 |
| Open LLM Leaderboard - GPQA | 19.69 | Score | 98.9 |
| EuroEval Finnish NLU - Tydiqa FI | 77.11 | Reading comprehension Score (%) | 98.8 |
| EuroEval Finnish NLU | 69.49 | NLU Average Score (%) | 98.4 |
| EuroEval French NLU - FQuAD | 76.09 | Reading comprehension Score (%) | 97.6 |
| EuroEval Portuguese NLU - MultiWikiQA PT | 78.57 | Reading comprehension Score (%) | 97.5 |
| Open Japanese LLM - Wikicorpus E TO J Bert Score JA F1 | 85.84 | Score (%) | 96.4 |
| EuroEval Italian NLU - MultiNERD IT | 84.06 | Named entity recognition Score (%) | 96.3 |
| Open Japanese LLM - Wikicorpus E TO J Comet Wmt22 | 85.54 | Score (%) | 95.9 |
Interactive version: theaggregate.ai/model?slug=llama-3-70b · How It Works · Data refreshed daily, snapshot 2026-09-05.