Meta
Llama 3.1-70B benchmark results
7 results on 7 benchmarks from 4 sources.
Open weights Released: Jul 23, 2024 Updated Oct 22, 2026
Also listed as: llama-3.1-70b-instruct · Llama-3.1-70B-Instruct · Llama 3.1 Instruct Turbo (70B) · Llama-3.1-70B · meta-llama/Llama-3.1-70B-Instruct · meta-llama/Llama-3.1-70B
All benchmark results
| Benchmarks | Score | Rank | Setting | Source type | Date |
|---|---|---|---|---|---|
| LMArena Text | 1293 | 287 | Overall, style control on; 55240 votes; 95% CI 1290–1297“llama-3.1-70b-instruct” | Official leaderboard | Oct 8, 2026 |
| Open LLM Leaderboard v2 | 43.4% | 62 | Average of 6 normalised benchmarks; bfloat16; 70.554B params“meta-llama/Llama-3.1-70B-Instruct” | Official leaderboard | Mar 13, 2025 |
| GPQA Diamond | 44.2% | 228 | Epoch AI run, accuracy | Independent run | Jan 27, 2025 |
| MMLU | 80.1% | 17 | Exact match; 5-shot | Official leaderboard via Epoch AI | Jul 23, 2024 |
| MMLU-Pro | 52.5% | 160 | Overall accuracy; data source: TIGER-Lab; 70B“Llama-3.1-70B” | Official leaderboard | Mar 11, 2026 |
| AIME (OTIS Mock AIME 2024–2025) | 3.6% | 241 | Epoch AI run, OTIS Mock AIME 2024–2025 | Independent run | Feb 25, 2025 |
| MATH Level 5 | 36.7% | 75 | Epoch AI run, accuracy | Independent run | Jan 27, 2025 |
FAQ
Where do these Llama 3.1-70B scores come from?
From Epoch AI Benchmarking Hub, LMArena leaderboard dataset, MMLU-Pro leaderboard (TIGER-Lab), and Hugging Face Open LLM Leaderboard v2 (archived). Each row links to the original leaderboard, dataset or report, and vendor-reported numbers are labelled.