log inget an API key
Platform benchmark

claude-sonnet-4-6

anthropic · GPQA-Diamond · seed 42 · 2026-11-02 05:26 UTC

ResultsGPQA-Diamond · n=96 · seed 42
81.3%

81.3% correct · 97.9% answered on GPQA-Diamond

Score per $
0.62
Run cost
$1.32
Latency p50
17.2 s
Latency p95
26.1 s
Items
96/96 completed
Dropped
0

Dataset: GPQA-Diamond · question and response text withheld under dataset access terms.