GUI Agent Task on Mind2Web (live)
50.8Completion RateSeeAct-V + GPT-4o + UGround
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| SeeAct-V + GPT-4o + UGroundInputs=Image (SeeAct-V), Planner=GPT-4o, Grounding=UGround2024.10 | 50.8 | 19.2 | |
| SeeAct-V + GPT-4 + UGroundInputs=Image (SeeAct-V), Planner=GPT-4, Grounding=UGround2024.10 | 50.7 | 23.1 | |
| GPT-4o + ChoiceInputs=Text, Planner=GPT-4o, Grounding=Choice2024.10 | 47.6 | 22.1 | |
| GPT-4 + ChoiceInputs=Text, Planner=GPT-4, Grounding=Choice2024.10 | 44.3 | 21.1 |