Mistral AI
Mixtral 8x22B benchmark results
8 results on 7 benchmarks from 5 sources.
Open weights Released: Apr 17, 2024 Updated Oct 22, 2026
Also listed as: mixtral-8x22b-instruct-v0.1 · open-mixtral-8x22b · Mixtral-8x22B-v0.1 · Mixtral (8x22B) · Mixtral-8x22B-Instruct-v0.1 · mistralai/Mixtral-8x22B-Instruct-v0.1 · mistralai/Mixtral-8x22B-v0.1
All benchmark results
| Benchmarks | Score | Rank | Setting | Source type | Date |
|---|---|---|---|---|---|
| LMArena Text | 1230 | 330 | Overall, style control on; 51416 votes; 95% CI 1225–1234“mixtral-8x22b-instruct-v0.1” | Official leaderboard | Oct 8, 2026 |
| Open LLM Leaderboard v2 | 33.9% | 731 | Average of 6 normalised benchmarks; bfloat16; 140.621B params“mistralai/Mixtral-8x22B-Instruct-v0.1” | Official leaderboard | Mar 13, 2025 |
| GPQA Diamond | 34.1% | 250 | Epoch AI run, accuracy | Independent run | Jan 27, 2025 |
| MMLU | 77.8% | 28 | Exact match; 5-shot | Official leaderboard via Epoch AI | Apr 17, 2024 |
| MMLU-Pro | 56.3% | 150 | Overall accuracy; data source: TIGER-Lab; 176B“Mixtral-8x22B-Instruct-v0.1” | Official leaderboard | Mar 11, 2026 |
| HumanEval / HumanEval+ | 76.2% | 28 | HumanEval (base tests), pass@1 greedy“Mixtral-8x22B-Instruct-v0.1” | Official leaderboard | Dec 26, 2024 |
| HumanEval / HumanEval+ | 72.0% | 23 | HumanEval+ (extended tests), pass@1 greedy“Mixtral-8x22B-Instruct-v0.1” | Official leaderboard | Dec 26, 2024 |
| MATH Level 5 | 24.2% | 82 | Epoch AI run, accuracy | Independent run | Jan 27, 2025 |
FAQ
Where do these Mixtral 8x22B scores come from?
From Epoch AI Benchmarking Hub, EvalPlus leaderboard, 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.