Disease prediction on Longitudinal medical record benchmark (test)
90.47AccuracyCOTCAgent
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| COTCAgentReader / Mode=Baichuan-M22026.05 | 90.47 | 86.74 | 92.18 | 82.91 | |
| DirPredReader / Mode=GPT-4o2026.05 | 88.27 | 85.06 | 83.98 | 70.87 | |
| o4-mini + toolsReader / Mode=OpenAI2026.05 | 87.51 | 84.12 | 89.82 | 80.74 | |
| Claude 3.7 SonnetReader / Mode=single-pass2026.05 | 86.94 | 83.55 | 89.38 | 80.16 | |
| Gemini 2.5 ProReader / Mode=single-pass2026.05 | 86.08 | 82.71 | 88.61 | 79.31 | |
| Qwen3-Next-80B-A3BReader / Mode=single-pass2026.05 | 85.14 | 81.87 | 87.51 | 77.62 | |
| KAREReader / Mode=GPT-4o2026.05 | 83.76 | 80.05 | 89.43 | 83.24 | |
| Google agt.Reader / Mode=GPT-4o2026.05 | 80.47 | 76.84 | 85.19 | 72.08 | |
| TimeCAPReader / Mode=GPT-4o2026.05 | 78.31 | 74.18 | 83.06 | 69.43 |