Object Hallucination Evaluation
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
92.37Random Accuracy
21
Apr 1, 2026
88.1F1 Score
21
Feb 26, 2026
90.4POPE Score
20
Jun 26, 2026
92.33Accuracy
18
Apr 23, 2026
56.2CHAIRs
18
Feb 26, 2026
88.9Accuracy
18
Feb 26, 2026
96.67Accuracy
18
Feb 26, 2026
85.9F1 Score
16
Jun 24, 2026
58.8CHAIRs Score
15
Feb 26, 2026
56.2CHAIR Score (Scene)
15
Feb 26, 2026
64.7Score
15
Feb 26, 2026
87Accuracy
13
Apr 21, 2026
42.53Accuracy
12
Jul 2, 2026
56Hallucination Rate (%)
12
May 7, 2026
92.93Accuracy (Random)
12
Apr 15, 2026
42.3Hallucination Score (Hal)
11
Jun 26, 2026
86.93Random F1 Score
10
Jul 7, 2026
88.5Accuracy
10
Jun 18, 2026
45.1CHAIR_s
10
Feb 26, 2026
88POPE Score (Zh)
10
Feb 26, 2026
89.38POPE Score
9
Jun 17, 2026
64Inference (ms)
9
May 22, 2026
33.9SHR
9
Feb 26, 2026
87.9Accuracy
8
Jun 4, 2026
12Total Time
8
Jun 2, 2026