xAI · model
Grok 2 mini benchmark scores
As of 2026-10-29, Grok 2 mini (xAI) has measured scores on 4 benchmarks and estimated scores on 31 more.
| Benchmark | Score | Source |
|---|---|---|
| AA-GPQA Diamond | 48.8% | estimated ± 6.3 pp, medium confidence |
| AA-HLE | 4.4% | estimated ± 5.0 pp, high confidence |
| CritPt | 0.0% | estimated ± 3.4 pp, medium confidence |
| GPQA Diamond | 51.4% | estimated ± 1.2 pp, high confidence |
| HLE | 2.3% | estimated ± 6.9 pp, low confidence |
| AA-MMMU-Pro | 49.3% | estimated ± 3.0 pp, high confidence |
| MMLU-Pro | 72.0% | measured |
| Artificial Analysis Intelligence Index | 10.8% | estimated ± 2.1 pp, high confidence |
| BBH | 91.2% | estimated ± 2.4 pp, low confidence |
| GPQA | 51.0% | measured |
| MMLU | 86.2% | measured |
| AA-Omniscience Accuracy | 18.3% | estimated ± 1.5 pp, high confidence |
| MMMLU | 73.2% | estimated ± 1.7 pp, high confidence |
| MMMU | 63.2% | measured |
| AA MMLU-Pro | 81.7% | estimated ± 1.0 pp, low confidence |
| HLE w/o tools | 11.6% | estimated ± 5.2 pp, low confidence |
| AI2D_TEST | 87.7% | estimated ± 2.0 pp, high confidence |
| C-Eval | 82.1% | estimated ± 1.2 pp, medium confidence |
| CountBench | 93.5% | estimated ± 3.2 pp, high confidence |
| DynaMath | 80.3% | estimated ± 5.8 pp, low confidence |
| ERQA | 47.1% | estimated ± 5.2 pp, low confidence |
| MathVision | 60.4% | estimated ± 3.6 pp, high confidence |
| MLVU (M-Avg) | 78.2% | estimated ± 2.4 pp, high confidence |
| MMLU-Redux | 85.5% | estimated ± 0.6 pp, high confidence |
| MMMU-Pro | 49.1% | estimated ± 2.3 pp, high confidence |
| RealWorldQA | 76.8% | estimated ± 3.6 pp, high confidence |
| ScreenSpot Pro | 53.0% | estimated ± 7.3 pp, low confidence |
| SimpleVQA | 46.5% | estimated ± 7.1 pp, medium confidence |
| SuperGPQA | 51.5% | estimated ± 12.1 pp, low confidence |
| Video-MME (w/o subtitle) | 66.9% | estimated ± 1.3 pp, high confidence |
| Video-MME (with subtitle) | 73.7% | estimated ± 3.2 pp, high confidence |
| VideoMMMU | 54.0% | estimated ± 3.8 pp, high confidence |
| We-Math | 71.4% | estimated ± 7.2 pp, low confidence |
| SimpleQA | 22.5% | estimated ± 13.1 pp, low confidence |
| MMVU | 62.1% | estimated ± 3.8 pp, medium confidence |