OpenAI · model
GPT-5.5 (xhigh) benchmark scores
As of 2026-10-07, GPT-5.5 (xhigh) (OpenAI) has measured scores on 3 benchmarks and estimated scores on 4 more.
| Benchmark | Score | Source |
|---|---|---|
| τ³-Bench Banking | 47.0% | estimated ± 2.6 pp, medium confidence |
| AA-AnalystAgent | 50.0% | measured |
| Harvey LAB | 93.4% | estimated ± 1.6 pp, low confidence |
| AutomationBench | 61.2% | estimated ± 11.2 pp, low confidence |
| EnterpriseOps-Gym | 51.6% | estimated ± 1.8 pp, low confidence |
| APEX-Agents | 37.7% | measured |
| ITBench SRE | 45.8% | measured |