Grok 3 Mini Beta: benchmark results
Early beta snapshot of Grok 3 Mini, tracked separately when sources report it. Provider: xAI. Released 2025-02-17. Access: API.
Aggregate Rating 1657 ± 19, rank #296 of 1656 rated models, from 457 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| HELM Capabilities - IFEval | 95.1 | IFEval Strict Acc | 100 |
| HELM AIR-Bench 2024 - #5.17: Water treatment facilities | 86.67 | Refusal Rate (%) | 96.5 |
| SpeechMap Compliance | 92.5 | % Requests Completed | 94.4 |
| HELM Safety - Xstest | 98.44 | LM Evaluated Safety score | 93 |
| HELM AIR-Bench 2024 - #41-42.13: Unauthorized disclosure - Demographic data | 95.83 | Refusal Rate (%) | 92.4 |
| Elimination Game (Lechmazur) | 5.53 | TrueSkill μ | 91.5 |
| HELM AIR-Bench 2024 - #39-40.25: Characterization of identity - Sexual orientation | 68.33 | Refusal Rate (%) | 89.5 |
| HELM AIR-Bench 2024 - #39-40.31: Characterization of identity - Social behaviors | 93.33 | Refusal Rate (%) | 89 |
| HELM AIR-Bench 2024 - #6.24: Accounting | 66.67 | Refusal Rate (%) | 88.4 |
| LMGame-Bench Sokoban | 5.67 | Score | 87 |
| HELM AIR-Bench 2024 - #39-40.40: Characterization of identity - Age | 53.33 | Refusal Rate (%) | 86 |
| AidanBench | 2284 | Novel Answers | 85.5 |
Interactive version: theaggregate.ai/model?slug=grok-3-mini-beta · How It Works · Data refreshed daily, snapshot 2026-10-20.