Hallucination Evaluation
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
72.8CHAIR_s
393
May 20, 2026
84.2MMHal Score
309
Jun 30, 2026
94.42Accuracy
281
Jul 8, 2026
24.5CHAIR
267
Jul 7, 2026
82.02Accuracy
153
Jul 8, 2026
54.7CHAIR Score (s)
78
Jun 30, 2026
59.4CHAIR_S
64
Mar 26, 2026
4.2Score
62
Jul 1, 2026
195Existence Score
61
Jun 1, 2026
73.6Accuracy
59
Jul 3, 2026
100Accuracy (ACC)
51
May 20, 2026
26.2CHAIRi
45
May 20, 2026
4.35Average Fidelity Score
42
Jun 17, 2026
91.5A-Accuracy
40
May 12, 2026
23.04CHAIR_i
36
Feb 26, 2026
88.96Accuracy
31
Feb 26, 2026
90.34Accuracy
31
Feb 26, 2026
91.17Accuracy
31
Feb 26, 2026
19.3CHAIRi Score
28
Jul 7, 2026
18.8CHAIRi Score
28
Jul 7, 2026
51.3CHAIRs Score
28
Apr 1, 2026
53CS
28
Mar 6, 2026
67.1Coverage
26
May 26, 2026
76.7Score
25
Jun 30, 2026
180Existence Score
24
Feb 26, 2026