List your company

Click-in-target on OOD Evaluation Suite (ScreenSpot-Pro, OSWorld-G, ScreenSpot-V2) (unweighted mean)

60Click-in-Target Accuracy (SS-Pro)Qwen3.5-4B (base)
47.10450.45253.857.148Jun 24, 2026

Evaluation results

MethodLinks
Qwen3.5-4B (base)Base model=Qwen3.5-4B, Cold-start regime=None, RL regime=None2026.066056.3—70
SFD-4B +RL EarlyBase model=Qwen3.5-2B, Cold-start regime=SFD-4B, RL regime=Early2026.0655.754.588.866.3
SFD-EMA +RL EarlyBase model=Qwen3.5-2B, Cold-start regime=SFD-EMA, RL regime=Early2026.0655.151.68965.2
SFD-4BBase model=Qwen3.5-2B, Cold-start regime=SFD-4B, RL regime=None2026.0654.750.889.565
SFD-EMABase model=Qwen3.5-2B, Cold-start regime=SFD-EMA, RL regime=None2026.0653.551.288.864.5
SFD-EMA +RL LateBase model=Qwen3.5-2B, Cold-start regime=SFD-EMA, RL regime=Late2026.0653.25185.463.2
SFD-4B +RL LateBase model=Qwen3.5-2B, Cold-start regime=SFD-4B, RL regime=Late2026.065351.488.964.4
SFT +RL EarlyBase model=Qwen3.5-2B, Cold-start regime=SFT, RL regime=Early2026.0652.752.487.764.3
Qwen3.5-2B (base)Base model=Qwen3.5-2B, Cold-start regime=None, RL regime=None2026.0652.247.58360.9
SFT +RL LateBase model=Qwen3.5-2B, Cold-start regime=SFT, RL regime=Late2026.065051.687.663.1
SFTBase model=Qwen3.5-2B, Cold-start regime=SFT, RL regime=None2026.0647.64985.960.8