Depression Detection on DAIC-WOZ (Accuracy and F1 Variants)
85.1AccuracyDep-LLM
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Dep-LLMFoundation LLM=Gemma3-12B-Instruct2026.06 | 85.1 | 81.8 | 84.9 | 74.1 | 89.6 | |
| Claude-Opus-4.6Model Category=Closed-Source Commercial LLMs2026.06 | 78.7 | 76.3 | 79.4 | 68.8 | 83.9 | |
| MentalBERTModel Category=Domain-specific LLMs2026.06 | 76.6 | 72.6 | 76.8 | 62.1 | 83.1 | |
| BioMistralModel Category=Domain-specific LLMs2026.06 | 74.5 | 73 | 75.5 | 66.7 | 79.3 | |
| MeditronModel Category=Domain-specific LLMs2026.06 | 74.5 | 69.5 | 74.5 | 57.1 | 81.8 | |
| Grok-4.3Model Category=Closed-Source Commercial LLMs2026.06 | 72.3 | 69.6 | 73.3 | 60.6 | 78.6 | |
| MentalRoBERTaModel Category=Domain-specific LLMs2026.06 | 70.2 | 69.9 | 71.2 | 66.7 | 73.1 | |
| ClinicalBERTModel Category=Domain-specific LLMs2026.06 | 70.2 | 67.7 | 71.4 | 58.8 | 76.7 | |
| GPT-5.5Model Category=Closed-Source Commercial LLMs2026.06 | 70.2 | 69.1 | 71.5 | 63.2 | 75 | |
| Gemini-3.1-ProModel Category=Closed-Source Commercial LLMs2026.06 | 70.2 | 66.8 | 71.1 | 56.3 | 77.4 | |
| DeepSeek-V4Model Category=Closed-Source Commercial LLMs2026.06 | 70.2 | 65.7 | 70.7 | 53.3 | 78.1 | |
| MentalAlpacaModel Category=Domain-specific LLMs2026.06 | 68.1 | 65 | 69.2 | 54.5 | 75.4 | |
| MentaLLaMAModel Category=Domain-specific LLMs2026.06 | 66 | 63.1 | 67.3 | 52.9 | 73.3 |