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

Compositional Vision-Language Reasoning on Winoground

89.5Text Score

MTurk Human

13.5833.295372.71Mar 30, 2023Sep 28, 2023Mar 28, 2024Sep 27, 2024Mar 28, 2025Sep 26, 2025Mar 28, 2026
Updated 19d ago

Evaluation Results

MethodLinks
2023.05
89.588.585.5
2023.11
89.588.585.5
2023.09
89.588.585.5
2026.03
89.588.585.5
2026.03
767266
2023.11
75.2568.7558.75
2026.03
74.57262.8
2026.03
74.573.564.2
2023.11
69.2546.2539.25
2023.09
69.2546.2539.25
2026.03
66.866.354.8
2026.03
62.861.849.5
2026.03
5561.347.5
2026.03
51.550.536.5
2023.11
51.4931.4927.5
2023.11
4742.230.5
2023.09
4742.230.5
2023.11
46.53828.75
2023.09
46.53828.75
2023.11
45.544.9943
2023.09
454543
2023.11
44.9922.7518.75
2023.09
442623.5
2023.05
42.519.7516
2024.12
40.517.512.8
2023.05
39.2517.7514.25
2023.05
381410.5
2023.05
37.7517.7514.5
2024.12
36.514.810.3
2024.12
35.81510.5
2024.12
35.515.512
2026.03
352222
2024.12
33.81410.8
2024.12
32.314.811.3
2024.12
3212.89.3
2024.12
31.317.410.5
2023.03
31.2510.58
2026.03
30.810.58
2023.05
30.7510.257.25
2023.11
30.7510.58
2024.12
30.5107.8
2023.03
3010.758.25
2023.03
3011.59.5
2024.12
29.510.58.5
2024.12
29.510.58
2023.05
29.252717.5
2023.03
28.59.57.25
2024.12
27.512.38.3
2023.05
26.7510.756.5
2023.05
252516.67
2023.11
252516.67
2024.12
2510.57.3
2026.03
25256.3
2023.05
24.523.513.75
2023.05
22.757.754.5
2023.11
1912.511.3
2024.12
18.511.36.5
2023.05
16.515.756.75