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

Action Prediction on AndroidControl High v2

68.59Pass@1 Step Accuracy

Mem-W-8B

35.944444.419752.89561.3703May 10, 2026
Updated 22d ago

Evaluation Results

MethodLinks
2026.05
68.5981.1671.380.488.9482.96
2026.05
64.3--70.6--
2026.05
63.0774.6270.8574.8783.9280.72
2026.05
61.3--69.4--
2026.05
61.0674.6261.0670.180.482.51
2026.05
59.374.3769.9671.3683.1780.72
2026.05
59.3--67.3--
2026.05
59.0574.6269.9670.182.1677.58
2026.05
57.5--63.3--
2026.05
57--66.3--
2026.05
54.869.3567.7166.177.8979.82
2026.05
52.8--58.8--
2026.05
51.2667.8464.1364.0774.6275.34
2026.05
49.363.5769.0663.375.8877.13
2026.05
45.4864.0759.6459.873.6273.99
2026.05
45.270.657.463.184.1772.65
2026.05
42.7161.0666.3743.4765.3362.78
2026.05
40.4559.0560.0952.7669.673.09
2026.05
40--54--
2026.05
39.7--56.3--
2026.05
37.7--43.7--
2026.05
37.267.8448.434974.8759.64