Biomedical Temporal Reasoning on ChronoTQA 1.0 (120-question stratified subsample)
100Cross-disease ComparisonGPT-4o-mini
Evaluation Results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| GPT-4o-miniAccess=Web chat, Retrieval=No retrieval, API=No API, System Prompts=No system prompts2026.05 | 100 | 58 | 82 | 0 | 80 | 75 | 52.9 | 77.8 | 24.8 | |
| GeminiAccess=Web chat, Retrieval=No retrieval, API=No API, System Prompts=No system prompts2026.05 | 100 | 83 | 75 | 12 | 100 | 100 | 64 | 100 | 36 | |
| ClaudeAccess=Web chat, Retrieval=No retrieval, API=No API, System Prompts=No system prompts2026.05 | 100 | 75 | 55 | 0 | 80 | 75 | 51 | 77.8 | 26.8 | |
| DeepSeek V3Access=Web chat, Retrieval=No retrieval, API=No API, System Prompts=No system prompts2026.05 | 91 | 67 | 36 | 6 | 80 | 75 | 45.1 | 77.8 | 32.7 |