GPAgentBench-2K - Missed Local Management Rate: leaderboard
Metric: Missed local management rate (%; locally manageable cases the agent referred unnecessarily; lower is better; zero-shot GP agent with six clinical actions over 2,428 physician-validated primary-care cases, GPT-4o-mini patient simulator, at most 10 turns). Source: arxiv.org. Saturation forecast: Around 2028. 21 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Sonnet 4.6 | 13.4 |
| 2 | MiniMax-M2.7 | 25 |
| 3 | Claude Opus 4.7 | 27.8 |
| 4 | Qwen 3.7 Max | 29.2 |
| 5 | Lingshu-7B | 31 |
| 6 | DeepSeek V4 Pro | 33.5 |
| 7 | MiMo-V2.5-Pro | 34.5 |
| 8 | Lingshu-32B | 35.2 |
| 9 | MedGemma-27B | 35.2 |
| 10 | GPT-5.4 | 38.4 |
| 11 | Gemini 3.1 Pro (Preview) | 41.5 |
| 12 | GLM-5.1 | 41.5 |
| 13 | DeepSeek V4 Flash | 45.1 |
| 14 | Kimi K2.6 | 71.8 |
Interactive version: theaggregate.ai/benchmark?slug=gpagentbench-2k-missed-local-management-rate · How It Works · Data refreshed daily, snapshot 2026-09-26.