Google · model
Gemini 3.8 Flash (high) benchmark scores
As of 2026-10-07, Gemini 3.8 Flash (high) (Google) has measured scores on 12 benchmarks and estimated scores on 3 more.
| Benchmark | Score | Source |
|---|---|---|
| Terminal-Bench 2.1 | 87.6% | measured |
| Terminal-Bench 4.0 | 19.7% | measured |
| Terminal-Bench Science 0.1 | 10.0% | measured |
| GPQA Diamond | 95.3% | measured |
| SciCode | 56.6% | measured |
| CritPt | 18.3% | measured |
| GDP.pdf | 21.0% | measured |
| τ³-Bench Banking | 44.9% | measured |
| MMMU-Pro | 86.0% | measured |
| AA-AnalystAgent | 47.8% | estimated ± 7.8 pp, medium confidence |
| Harvey LAB | 93.8% | estimated ± 1.6 pp, low confidence |
| Humanity's Last Exam | 47.8% | measured |
| AutomationBench | 59.9% | measured |
| EnterpriseOps-Gym | 41.9% | estimated ± 5.3 pp, medium confidence |
| ITBench SRE | 52.5% | measured |