Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

GUI Grounding on MMBench-GUI L2 (in-domain)

58Accuracy

GRPO w. Max.

44.4847.9951.555.01Nov 11, 2025
Updated 19d ago

Evaluation Results

MethodLinks
2025.11
58
2025.11
57.6
2025.11
56
2025.11
55
2025.11
55
2025.11
54.6
2025.11
53.8
2025.11
53.6
2025.11
51
2025.11
49.6
2025.11
49.4
2025.11
47
2025.11
45