Safety Evaluation
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
97.5Safe@1
34
May 21, 2026
54.5Reasoning Success Rate (FFR)
32
May 7, 2026
100Safe Response Rate
32
Feb 26, 2026
98.4Safety Score
32
Jun 1, 2026
0ASR
31
Jun 9, 2026
96Accuracy
30
Jun 11, 2026
100Overall Safety Score
30
May 7, 2026
18ASR (Hate Speech)
30
Feb 26, 2026
2.68Hp Score
28
Jun 1, 2026
89Accuracy
28
May 11, 2026
1.22Harmfulness
28
May 11, 2026
16Harm-1 Score
28
Apr 8, 2026
68.44HarmBench Score
28
Feb 26, 2026
0.23ASR
27
Apr 28, 2026
92.73Safety Score
27
Feb 26, 2026
92.1Score
27
Feb 26, 2026
0.72ASR
27
Apr 28, 2026
0.08Wildguard Test Score
27
Jun 9, 2026
73ASR (OM)
26
Jun 4, 2026
0ASR (OM)
26
Jun 4, 2026
0.62ASR (OM)
26
Jun 4, 2026
0ASR (OM)
26
Jun 4, 2026
69.4Safety
26
Mar 18, 2026
91Safety Score (DAN)
26
Jun 9, 2026
99.9Refusal Rate
25
Jul 1, 2026