Image Captioning Evaluation
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
66.2Kendall-c Tau_c
161
Jun 30, 2026
56.6Kendall-b Correlation (tau_b)
145
Jun 30, 2026
60.8Kendall Tau-c (tau_c)
114
Jun 30, 2026
56.4Kendall tau_c
76
Feb 26, 2026
57.5Kendall tau_c
66
Jun 30, 2026
70HC
66
Feb 26, 2026
38.8Kendall tau_b
65
Feb 26, 2026
87.8Accuracy
44
May 26, 2026
76.57Human Score
40
Apr 7, 2026
99.2Accuracy (4-ref)
33
Jun 30, 2026
89.96GPT Score
32
Jun 26, 2026
26.2tau_c
18
Feb 26, 2026
64.2Kendall's tau-c
17
Feb 26, 2026
31Precision
16
Feb 26, 2026
34Precision
16
Feb 26, 2026
42.3Fluency
15
Apr 7, 2026
51.3Fluency Score
15
Apr 7, 2026
87.8Real Style Score
12
Feb 26, 2026
62.42Accuracy
7
Feb 26, 2026
76.9Mean Accuracy
4
Feb 26, 2026