Llama 3.3 70B Instruct — benchmark results
Meta Llama 3.3 70B instruction-tuned checkpoint. Provider: Meta. Released 2024-12-06. Access: Open.
Unified ELO 1543 ± 5, rank #642 of 1776 rated models, from 1110 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| BlueBench - Product Help | 86.64 | Score (%) | 100 |
| BlueBench - Summarization | 19.47 | Score (%) | 100 |
| DiagFlowBench | 85 | Step Accuracy (self-reported) | 100 |
| EVALITA - evalita NER | 44.58 | CPS | 100 |
| EVALITA - hate-speech-detection | 77.82 | CPS | 100 |
| EuroEval Estonian Summarization - ERR News | 32.83 | Score (%) | 100 |
| EuroEval Norwegian Summarization - NO Sammendrag | 32.63 | Score (%) | 100 |
| French LLM Leaderboard - IFEval FR | 74.51 | Score (%) | 100 |
| HELM SeaHELM - LINDSEA Scalar Implicatures (id) | 92.31 | EM | 100 |
| Medical Information Response Audit (MIRA) | 2.95 | Underinformative simplification (D3) (self-reported) | 100 |
| Open LLM Leaderboard - IFEval | 89.98 | Score | 100 |
| Open Persian LLM - AUT Multiple Choice Persian | 71.4 | Accuracy (%) | 100 |
Interactive version: theaggregate.ai/model?slug=llama-3-3-70b-instruct · How the rankings work · Data refreshed daily, snapshot 2026-07-22.