Preference Reasoning (Recommendation) on SCREEN 1.0 (test) benchmark leaderboard · SOTA2 Research

Preference Reasoning (Recommendation) on SCREEN 1.0 (test)

39.41R@1SiPeR

Evaluation Results

Method	Links
SiPeR2026.04		39.41	54.95	63.8	50.36	51.95
GPT-4o2026.04		35.06	49.94	60.16	44.58	45.96
SiPeR w/o BI-INF2026.04		33.96	48.49	51.96	46.24	47.92
GPT-4o2026.04		33.45	49.32	58.15	42.21	44.58
Qwen2.5-VL2026.04		32.06	35.02	37.26	34.01	35.32
ReGeS2026.04		31.42	49.85	59.24	39.88	41.75
SiPeR w/o STE2026.04		30.26	43.88	51.16	40.71	42.54
LLaVA-NeXT2026.04		24.63	28.49	30.46	27.98	29.11
Qwen2.5-VL2026.04		21.18	23.24	23.96	22.95	23.52
Qwen2.5-VL2026.04		21.05	23.68	24.12	23.01	23.42
LLaVA-NeXT2026.04		16.71	18.22	18.8	16.67	17.2
LLaVA-NeXT2026.04		15.42	16.85	18.21	15.68	15.98
ALBEF2026.04		8.51	9.98	12.64	10.75	12.03