Multimodal Report Generation on DeepResearch Bench
3.72DLBPTAH
Evaluation Results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| PTAHGeneration Category=Multimodal Generation2026.05 | 3.72 | 4.42 | 4.79 | 4.35 | 4.01 | 4.39 | 3.78 | 3.61 | 3.74 | 3.71 | |
| QwQ-32BGeneration Category=Direct Generation2026.05 | 3.65 | — | — | — | — | — | 3.67 | 2.84 | 3.65 | 3.45 | |
| Qwen3-32BGeneration Category=Direct Generation2026.05 | 3.55 | — | — | — | — | — | 3.6 | 2.73 | 3.53 | 3.35 | |
| Search-o1Generation Category=Text-Only Generation2026.05 | 3.49 | — | — | — | — | — | 3.14 | 2.62 | 3.24 | 3.12 | |
| WebThinkerGeneration Category=Text-Only Generation2026.05 | 3.42 | — | — | — | — | — | 2.89 | 2.78 | 3.35 | 3.11 | |
| ReActGeneration Category=Text-Only Generation2026.05 | 3.13 | — | — | — | — | — | 3.21 | 2.41 | 3.55 | 3.08 | |
| LLM-IGeneration Category=Multimodal Generation2026.05 | 3.12 | 2.1 | 2.28 | 1.96 | 1.52 | 1.97 | 2.51 | 3.25 | 3.11 | 3 |