AutoMedBench - Lesion Detection: leaderboard
The lesion detection track of AutoMedBench (2026), which asks an LLM agent to complete end-to-end medical-AI research tasks built from public challenge datasets: four lesion detection tasks. Each run is scored 0 to 100 as the mean of the held-out result (mAP at IoU 0.5) and a five-stage workflow score (plan, setup, validate, inference, submit), averaged over the Lite and Standard tiers. Higher is better.
Metric: Overall run score (0-100) on the lesion detection track (4 detection tasks): the mean of the held-out task score (mAP at IoU 0.5, scaled to 0-100) and the workflow agentic score (Plan, Setup, Validate, Inference and Submit stages weighted 25/15/35/15/10, LLM judge for the first three); each model is the single agent of the benchmark harness (shared code-execution interface, system prompt, tool schema and default decoding; no vendor agent framework), runs averaged over the Lite and Standard tiers; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.6 | 85.7 | #102 |
| 2 | GLM-5 | 83.3 | #260 |
| 3 | Qwen 3.5 397B A17B | 81.4 | #246 |
| 4 | MiniMax-M2.5 | 80 | #582 |
| 5 | Gemini 3.1 Pro (Preview) | 77.2 | #94 |
| 6 | GPT-5.4 | 73.8 | #123 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=automedbench-lesion-detection · How It Works · Data refreshed daily, snapshot 2026-11-13.