Safety Evaluation
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
4.6ASR (WJ)
7
Apr 23, 2026
0Refusal Rate
7
Mar 5, 2026
95.4Task Performance
7
Feb 26, 2026
0.2ASR
7
Feb 26, 2026
1ASR
7
Feb 26, 2026
78.5ASR
7
Feb 26, 2026
93.4Safety Score (Avg)
7
Feb 26, 2026
29.1ASR (Hate Speech)
7
Feb 26, 2026
20.65Porn
7
Feb 26, 2026
95.9Safety Score
7
Feb 26, 2026
99.7Safety Score
7
Feb 26, 2026
100Safety Score
7
Feb 26, 2026
7.89Racial Bias Score
7
Feb 26, 2026
11.2SafeBench Score
7
Feb 26, 2026
10.5ASR
6
Jul 7, 2026
10.8ASR
6
Jul 7, 2026
62ASR
6
Jul 7, 2026
93Refusal Rate
6
Jun 26, 2026
0Refusal Rate
6
Jun 26, 2026
54.9Refusal Rate
6
Jun 26, 2026
30.7Refusal Rate
6
Jun 26, 2026
51.6Refusal Rate
6
Jun 26, 2026
-0.153Unsafe Score
6
Jun 25, 2026
57.5Offensive Attack Success Rate
6
Jun 25, 2026
3.12Harm Score
6
Jun 16, 2026