Table 1:

Experimental results on the CrossDocked dataset. The best results are shown in bold, and the second-best results are underlined.

Model	Vina Dock (↓)		High Affinity (↑)		QED (↑)		SA (↑)		Diversity (↑)		Size
	Avg.	Med.	Avg.	Med.	Avg.	Med.	Avg.	Med.	Avg.	Med.	Avg.	Med.
Reference	−7.117	−6.905	-	-	0.476	0.468	0.728	0.740	-	-	22.75	21.50
FDA drugs	−7.027	−7.169	0.440	0.380	0.567	0.564	0.760	0.758	0.792		24.47
ZINC 30 atoms	−7.855	−8.088	0.607	0.735	0.576	0.577	0.737	0.736	0.655		30.0
CVAE	−6.114	−6.118	0.103	0.026	0.390	0.419	0.591	0.580	0.655	0.666	19.97	20.19
AR	−6.751	−6.707	0.459	0.340	0.505	0.499	0.635	0.634	0.698	0.703	17.78	17.54
Pocket2Mol	−7.200	−6.815	0.601	0.593	0.574	0.579	0.754	0.760	0.741	0.781	17.84	16.53
TamGen	−7.475	−7.775	0.526	0.645	0.559	0.559	0.771	0.759	0.747	0.745	23.13	23.29
TargetDiff	−7.788	−7.964	0.683	0.634	0.474	0.485	0.584	0.571	0.717	0.714	24.44	24.64
Gemma-3–27B	−7.050	−7.024	0.416	0.281	0.700	0.711	0.860	0.868	0.757	0.765	19.34	18.94
GPT 4o	−7.198	−7.257	0.432	0.294	0.789	0.803	0.870	0.878	0.767	0.767	19.70	19.64
DeepSeekV3	−7.230	−7.170	0.443	0.241	0.743	0.756	0.855	0.867	0.771	0.772	18.96	19.00
Assay2Mol (Gemma-3–27B)	−8.064	−8.280	0.610	0.732	0.585	0.606	0.821	0.834	0.742	0.611	26.59	26.65
Assay2Mol (GPT 4o)	−7.796	−7.881	0.548	0.576	0.600	0.630	0.790	0.801	0.542	0.547	25.90	25.59
Assay2Mol (DeepSeekV3)	−7.861	−7.936	0.557	0.634	0.616	0.647	0.813	0.820	0.593	0.608	24.46	24.26
Assay2Mol (DeepSeekV3 <30%)	−7.826	−7.925	0.562	0.673	0.594	0.619	0.814	0.820	0.609	0.628	24.49	24.73