Dialogue reasoning on IndiRef
50Temporal AccuracyAgentic-Image
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Agentic-Image2026.04 | 50 | 24 | 44 | 58 | |
| Agentic-Text2026.04 | 42 | 26 | 44 | 46 | |
| Qwen-QwQFramework=Full Dialog (FD)2026.04 | 32 | 38 | 40 | 40 | |
| Qwen3-VL-ThinkingFramework=Full Dialog (FD)2026.04 | 18 | 10 | 20 | 24 |