Meta
Llama 2-70B benchmark results
9 results on 7 benchmarks from 4 sources.
Open weights Released: Jul 18, 2023 Updated Oct 22, 2026
Also listed as: llama-2-70b-chat · Llama-2-70b-chat-hf · Llama-2-70b-hf · Llama 2 (70B) · LLAMA 2 70B · LLaMA 2 70B · LLaMA 2 70B · Llama-2-70b-chat · LLaMA2-Chat 70B · Llama-2-70B
All benchmark results
| Benchmarks | Score | Rank | Setting | Source type | Date |
|---|---|---|---|---|---|
| LMArena Text | 1171 | 365 | Overall, style control on; 38492 votes; 95% CI 1165–1176“llama-2-70b-chat” | Official leaderboard | Oct 8, 2026 |
| Open LLM Leaderboard v2 | 18.4% | 2905 | Average of 6 normalised benchmarks; float16; 68.977B params“meta-llama/Llama-2-70b-hf” | Official leaderboard | Mar 13, 2025 |
| GPQA Diamond | 26.3% | 271 | Epoch AI run, accuracy | Independent run | Jan 27, 2025 |
| MMLU | 68.9% | 27 | Exact match; 5-shot | Paper / tech report via Epoch AI | Jul 18, 2023 |
| MMLU | 59.4% | 51 | Exact match; 0-shot | Paper / tech report via Epoch AI | Jul 18, 2023 |
| MMLU | 69.5% | 54 | Exact match; 5-shot | Official leaderboard via Epoch AI | Jul 18, 2023 |
| MMLU-Pro | 37.5% | 199 | Overall accuracy; data source: TIGER-Lab; 70B“Llama-2-70B” | Official leaderboard | Mar 11, 2026 |
| AIME (OTIS Mock AIME 2024–2025) | 0.0% | 265 | Epoch AI run, OTIS Mock AIME 2024–2025 | Independent run | Feb 25, 2025 |
| MATH Level 5 | 3.3% | 104 | Epoch AI run, accuracy | Independent run | Jan 27, 2025 |
FAQ
Where do these Llama 2-70B 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.