List your company
ResearchFormerly WizWand

Scientific Question Answering on SuperGPQA*

62.4AccuracyGPT-5
39.62445.53751.4557.363Aug 26, 2025
Updated 5mo ago

Evaluation results

MethodLinks
GPT-5Reasoning Effort=High2025.0862.43.8
Gemini-2.5-ProReasoning Effort=High2025.0860.40.3
Gemini-2.5-ProReasoning Effort=Low2025.0860.1—
o3Reasoning Effort=High2025.0859.54.6
GPT-5Reasoning Effort=Low2025.0858.6—
o4-miniReasoning Effort=High2025.0857.18.5
o3Reasoning Effort=Low2025.0854.9—
o3-miniReasoning Effort=High2025.085413.5
Claude-Sonnet-4Reasoning Effort=High2025.0849.84.6
o4-miniReasoning Effort=Low2025.0848.6—
Claude-Sonnet-4Reasoning Effort=Low2025.0845.2—
o3-miniReasoning Effort=Low2025.0840.5—