Mixtral 8x22B Instruct: benchmark results
Mistral's Apache-2.0 sparse-MoE instruct model (141B total/39B active, 64K context), strong at math and coding among April 2024 open models. Provider: Mistral. Released 2024-04-17. Access: Open.
Aggregate Rating 1516 ± 15, rank #728 of 1656 rated models, from 539 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| IKP - T2 | 100 | Accuracy (%) | 92.6 |
| ENAMEL | 40.8 | Greedy eff@1 (%) (imported) | 89.7 |
| AGC-Bench - grapheval_ai_researcher | 0.81 | Dataset z-score | 87.7 |
| HELM AIR-Bench 2024 - #4.3: Adversely affects legal rights | 80 | Refusal Rate (%) | 86.6 |
| AGC-Bench - tinystories | 0.94 | Dataset z-score | 85.4 |
| AGC-Bench - story_quality | 0.98 | Dataset z-score | 84.6 |
| HELM TORR - Numeric Nlg | 19.14 | ROUGE-L | 84.6 |
| HELM TORR - Scigen | 16.64 | ROUGE-L | 84.6 |
| AGC-Bench - future_ideas | 0.88 | Dataset z-score | 79.3 |
| HELM AIR-Bench 2024 - #39-40.51: Classification of individuals - Social behaviors | 96.67 | Refusal Rate (%) | 79.1 |
| AGC-Bench - schnovel | 0.9 | Dataset z-score | 78.7 |
| HELM AIR-Bench 2024 - #14.31: Age | 100 | Refusal Rate (%) | 77.3 |
Interactive version: theaggregate.ai/model?slug=mixtral-8x22b-instruct · How It Works · Data refreshed daily, snapshot 2026-10-20.