evals.report
BenchmarksLabsCompareRun guidesIn the wild

GPT-6 Luna

OpenAI · GPT. Released Sep 22, 2026.

GPT-6 Luna is a model from OpenAI in the GPT family, released Sep 22, 2026. evals.report tracks 5 reported GPT-6 Luna benchmark scores across FrontierCode, AA-Omniscience: Knowledge and Hallucination Benchmark, ARC-AGI-1, ARC-AGI-2, DeepSWE — each shown with its benchmark, metric, source status, and date, and never combined into a single ranking.

5 results

Benchmark results 5

Compare this model
BenchmarkCategoryScoreMetricStatusDate
FrontierCodeCoding42.4%weighted score (Main)OfficialSep 22, 2026Details
AA-Omniscience: Knowledge and Hallucination BenchmarkReasoning1AA-Omniscience IndexOfficialSep 22, 2026Details
ARC-AGI-1Reasoning86.67%accuracyOfficialSep 22, 2026Details
ARC-AGI-2Reasoning59.31%accuracyOfficialSep 22, 2026Details
DeepSWECoding66.6%% resolvedVerifiedSep 22, 2026Details