Llama 3.1 405B Instruct: benchmark results
Meta Llama 3.1 405B instruction-tuned checkpoint. Provider: Meta. Released 2024-07-23. Access: Open.
Aggregate Rating 1596 ± 7, rank #457 of 1656 rated models, from 1032 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AbstentionBench - underspecified context - ALCUNA - Precision | 93.27 | Abstention Precision (%) | 100 |
| BenCzechMark - NER | 94.81 | Duel Win Score (%) | 100 |
| BenCzechMark - NLI | 82.22 | Duel Win Score (%) | 100 |
| CRMArena - KQA (Function Calling Prompt) | 47.6 | F1 (%) | 100 |
| CRMArena - MTA (Function Calling Prompt) | 86.9 | Exact match (%) | 100 |
| CRMArena - NCR (ReAct) | 81.5 | Exact match (%) | 100 |
| GSM-8K (CoT) | 96.8 | Score (%) | 100 |
| Gorilla Benchmark API Bench | 35.3 | Score (%) | 100 |
| HELM MMLU - Nutrition | 92.81 | EM | 100 |
| HELM SeaHELM - XNLI (th) | 70 | EM | 100 |
| HELM ThaiExam - TPAT-1 | 68.1 | EM | 100 |
| MMLU (CoT) | 88.6 | Score (%) | 100 |
Interactive version: theaggregate.ai/model?slug=llama-3-1-405b-instruct · How It Works · Data refreshed daily, snapshot 2026-10-20.