Platform benchmark

gpt-5.6-luna

openai · GPQA-Diamond · seed 42 · 2026-07-15 18:56 UTC

ResultsGPQA-Diamond · n=96 · seed 42
89.6%

89.6% correct · 100.0% answered on GPQA-Diamond

Score per $
1.09
Run cost
$0.82
Latency p50
6.1 s
Latency p95
33.1 s
Items
96/96 completed
Dropped
0