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

General Multimodal Reasoning on LogicVista

50.56Score

Qwen2.5-VL-7B + SRPO

39.84842.62945.4148.191May 8, 2026
Updated 23d ago

Evaluation Results

MethodLinks
2026.05
50.56
2026.05
47.65
2026.05
47.42
2026.05
47.42
2026.05
46.97
2026.05
46.76
2026.05
46.53
2026.05
46.53
2026.05
46.53
2026.05
46.3
2026.05
46.08
2026.05
45.86
2026.05
45.63
2026.05
43.84
2026.05
43.62
2026.05
43.62
2026.05
42.95
2026.05
42.95
2026.05
41.16
2026.05
40.49
2026.05
40.26