Action Prediction on ScreenSpot Mobile
78Action Success Rate (Text)SeeClick
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| SeeClickModel Size=9.6B2026.05 | 78 | 52 | |
| CogAgentModel Size=18B2026.05 | 67 | 24 | |
| Skill-CMIB2026.05 | 64.8 | 65.1 | |
| Qwen2.5-VL-7B-Inst.Model Size=7B2026.05 | 60.8 | 53.3 | |
| FuyuModel Size=8B2026.05 | 41 | 1.3 | |
| GPT-4V2026.05 | 22.6 | 24.5 | |
| Qwen-VLModel Size=9.6B2026.05 | 9.5 | 4.8 | |
| MiniGPT-v2Model Size=7B2026.05 | 8.4 | 6.6 |