benchgap
OpenAI · model

GPT-6 Luna (Decisions) benchmark scores

As of 2026-11-18, GPT-6 Luna (Decisions) (OpenAI) has measured scores on 1 benchmark and estimated scores on 10 more.

BenchmarkScoreSource
Decision Index Full Score51.4%estimated ± 2.5 pp, high confidence
Decision Index Arts44.7%estimated ± 4.5 pp, high confidence
Decision Index Knowledge34.9%estimated ± 3.5 pp, high confidence
Decision Index Language51.7%estimated ± 5.6 pp, medium confidence
Decision Index Retrieval59.8%estimated ± 2.6 pp, high confidence
Decision Index Tools73.6%estimated ± 4.7 pp, high confidence
HLE-MultiChoice (Decision Index)11.5%estimated ± 2.1 pp, high confidence
Artificial Analysis Decision Index v1.051.3%estimated ± 4.0 pp, high confidence
Artificial Analysis Decision Index60.2%measured
When2Call (Decision Index)75.5%estimated ± 5.5 pp, medium confidence
WildGuardTest (Decision Index)89.9%estimated ± 2.2 pp, high confidence