Llama 3 8B: benchmark results
Meta's Llama 3 8B, the base pretrained checkpoint in the Llama 3 series, released under the Llama 3 community license. Provider: Meta. Released 2024-04-18. Access: Open.
Aggregate Rating 1395 ± 8, rank #1242 of 1656 rated models, from 397 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| EuroEval Portuguese NLU - MultiWikiQA PT | 77.31 | Character F1 (%) | 93.8 |
| EuroEval Finnish NLU - Tydiqa FI | 74.51 | Character F1 (%) | 92.5 |
| EuroEval Dutch NLU - DBRD | 91.87 | Matthews correlation coefficient (%) | 90.2 |
| EuroEval Finnish NLU - Scandisent FI | 92.2 | Matthews correlation coefficient (%) | 88.3 |
| EuroEval French NLU - FQuAD | 72.66 | Character F1 (%) | 87.4 |
| EuroEval Spanish NLU - MLQA ES | 65.7 | Character F1 (%) | 85.8 |
| IFEval Leaderboard | 78.84 | Final Score | 80 |
| EuroEval Portuguese NLU - SST-2 PT | 82.5 | Matthews correlation coefficient (%) | 79.3 |
| European LLM Leaderboard - FLORES200 Target | 15.84 | FLORES200 Target Score | 79.2 |
| European LLM Leaderboard - Zero-Shot Accuracy | 50.64 | Average Accuracy (%) | 78 |
| European LLM Leaderboard - FLORES200 | 15.57 | Average FLORES200 Score | 77.1 |
| European LLM Leaderboard - FLORES200 Source | 15.3 | FLORES200 Source Score | 75.6 |
Interactive version: theaggregate.ai/model?slug=llama-3-8b · How It Works · Data refreshed daily, snapshot 2026-10-20.