Trajectory Prediction on EHRWorld 110K
80.6Numerical Score @25 (Next Step)EHRWorld-14B
Evaluation Results
| Method | Links | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| EHRWorld-14BParameters=14B2026.02 | 80.6 | 71.6 | 88.8 | 78.4 | 66.7 | 85.1 | 92.8 | 91.3 | 98.4 | 83.9 | 76.5 | 92.6 | |
| GPT-5.2Version=5.22026.02 | 78.9 | 67.4 | 85.4 | 74.1 | 62.7 | 84.6 | 62 | 55.3 | 86.2 | 71.7 | 61.8 | 86.2 | |
| GLM-4.7Version=4.72026.02 | 77.5 | 66.3 | 85.5 | 67 | 47.6 | 71 | 42.6 | 30.5 | 71.6 | 62.4 | 48.1 | 77.1 | |
| DeepSeek-V3.2Version=3.22026.02 | 77 | 64.9 | 82.3 | 72.2 | 55.4 | 76.7 | 60.8 | 44.9 | 73.8 | 70 | 55.1 | 77.6 | |
| Qwen3-235B-A22B-InstructParameters=235B, Architecture=A22B, Type=Instruct2026.02 | 76.9 | 64.7 | 84.1 | 70.7 | 46.9 | 66.3 | 55.7 | 40.2 | 72.2 | 67.8 | 50.6 | 71.9 | |
| Qwen3-Next-80B-A3B-InstructParameters=80B, Architecture=A3B, Type=Instruct2026.02 | 75.9 | 63 | 83 | 71.5 | 53.8 | 75.2 | 55.8 | 44.2 | 79.2 | 67.7 | 53.7 | 76.3 | |
| Baichuan-M2-32BParameters=32B2026.02 | 75.2 | 60.1 | 79.9 | 69.2 | 49.9 | 72.1 | 52.7 | 41.8 | 79.3 | 65.7 | 50.6 | 73.4 |