HealthAgentBench - CT Abnormality Classification: leaderboard

Metric: 3D chest CT abnormality classification (10 tasks; success when every abnormality label is correct): task success rate (%) over three attempts per task, a trial succeeding only when it meets the category success criterion; terminal environments in the Harbor framework, xhigh reasoning effort, web browsing disabled. Source: arxiv.org. Saturation forecast: Around 2030. 10 models tracked.

Top models

#ModelScore
1GPT-5.5 (xHigh)33
2GPT-5.4 (xHigh)30
3GPT-5.4 Mini (xHigh)23
4Claude Opus 4.8 (xHigh)17
5Claude Opus 4.7 (xHigh)13
6GPT-5.3 Codex (xHigh)10

Interactive version: theaggregate.ai/benchmark?slug=healthagentbench-ct-abnormality-classification · How It Works · Data refreshed daily, snapshot 2026-09-29.