Multimodal Agent Task on TIR-Bench
47.75Average@4XSkill
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| XSkillModel=Gemini-3-Flash2026.03 | 47.75 | 75 | |
| Agent-KBModel=Gemini-3-Flash2026.03 | 36.62 | 62 | |
| AWMModel=Gemini-3-Flash2026.03 | 34.25 | 62.5 | |
| DCModel=Gemini-3-Flash2026.03 | 33.75 | 59 | |
| XSkillModel=Gemini-2.5-Pro2026.03 | 33.12 | 58 | |
| w/ ToolsModel=Gemini-3-Flash2026.03 | 32.37 | 58.5 | |
| XSkillModel=o4-mini2026.03 | 30.25 | 57.5 | |
| AWMModel=Gemini-2.5-Pro2026.03 | 29.75 | 53.5 | |
| Agent-KBModel=Gemini-2.5-Pro2026.03 | 29.13 | 52.5 | |
| No ToolsModel=Gemini-3-Flash2026.03 | 28.5 | 54 | |
| w/ ToolsModel=Gemini-2.5-Pro2026.03 | 28.38 | 54 | |
| XSkillModel=GPT-5-mini2026.03 | 28.25 | 56 | |
| DCModel=Gemini-2.5-Pro2026.03 | 27.62 | 51 | |
| DCModel=o4-mini2026.03 | 25.5 | 52 | |
| AWMModel=o4-mini2026.03 | 25.25 | 51 | |
| Agent-KBModel=GPT-5-mini2026.03 | 25.13 | 53 | |
| Agent-KBModel=o4-mini2026.03 | 24.75 | 52.5 | |
| w/ ToolsModel=o4-mini2026.03 | 24.62 | 49.5 | |
| DCModel=GPT-5-mini2026.03 | 24.5 | 53.5 | |
| AWMModel=GPT-5-mini2026.03 | 24.25 | 53 | |
| w/ ToolsModel=GPT-5-mini2026.03 | 23.5 | 50.5 | |
| No ToolsModel=Gemini-2.5-Pro2026.03 | 21.37 | 40 | |
| No ToolsModel=o4-mini2026.03 | 20.13 | 45 | |
| No ToolsModel=GPT-5-mini2026.03 | 20 | 46.5 |