ScreenSpot Pro
ScreenSpot Pro - visual agent benchmark for GUI screen element grounding and interaction.
Category
General Agents
Max Score
—
Score Type
percent
Active
Yes
Model Scores
Scores for ScreenSpot Pro
| Model | Score | Date | Verified |
|---|---|---|---|
| Qwen3.5-4B |
60.40%
|
01.02.2026 | Verified |
| Gemma4-31B-it |
99.50%
|
01.08.2026 | Unverified |
| Qwen3.6-27B |
100.00%
|
01.08.2026 | Unverified |
| Qwen3.5-122B-A10B |
85.71%
|
01.02.2026 | Verified |
| Qwen3.5-27B |
85.46%
|
01.02.2026 | Unverified |
| Claude-Sonnet-4.5 |
36.20
|
01.02.2026 | Unverified |
| Qwen3.5-35B-A3B |
81.20%
|
01.02.2026 | Unverified |
| Qwen3-VL-235B-A22B |
64.66%
|
01.02.2026 | Verified |
| Muse-Glimmer-30B |
98.25%
|
01.08.2026 | Verified |
| Qwen3-235B-A22B |
64.66%
|
01.02.2026 | Unverified |