Mental health clinical reasoning on MentalDx Bench (test)
83.99CAMENTALSEEK-DX-14B
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| MENTALSEEK-DX-14BModel=MENTALSEEK-DX-14B, Base Model=Qwen2.5-14B2026.02 | 83.99 | 70.08 | 69.38 | |
| MENTALSEEK-DX-7BModel=MENTALSEEK-DX-7B, Base Model=Qwen2.5-7B2026.02 | 82.58 | 67.56 | 67.13 | |
| Claude-Sonnet-4.5Model=Claude-Sonnet-4.52026.02 | 80.76 | 59.69 | 58.15 | |
| DeepSeek-R1Model=DeepSeek-R12026.02 | 78.37 | 55.2 | 54.49 | |
| QWQ-32BModel=QWQ-32B2026.02 | 76.26 | 47.89 | 46.63 | |
| GPT-OSS-120BModel=GPT-OSS-120B2026.02 | 75.84 | 52.53 | 50 | |
| Baichuan-M2Model=Baichuan-M22026.02 | 64.61 | 44.38 | 41.71 | |
| Qwen3-32BModel=Qwen3-32B2026.02 | 58.71 | 31.04 | 26.26 |