List your company

Real World Tasks on HealthBench Hard

38.6ScoreGPT-5-mini High
9.79217.27124.7532.229Jun 30, 2026

Evaluation results

MethodLinks
GPT-5-mini High2026.0638.6
GPT-5.2 High2026.0636.6
Seed2.0 Pro2026.0628.3
Gemini-3-Flash High2026.0621.5
Seed2.0 Lite2026.0620
Gemini-3-pro High2026.0615
Claude-Opus-4.52026.0611
Claude-Sonnet-4.52026.0610.9