MM-SafetyBench
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
MM-SafetyBench
95.3Text-only Recall
6
MM-SafetyBench QR (OOD)
0ASR
6
MM-SafetyBench WR (OOD)
0ASR (%)
6
MM-SafetyBench Turn 2
24.42ASR
5
MM-SafetyBench SD 1.0
48.3Illegal Activity Score
5
MM-SafetyBench
91.61Safety
4
MM-SafetyBench H
1Safety Score
4
MM-SafetyBench Claude-4.5 (ALL)
2.53Average Toxicity
3
MM-SafetyBench Gemini-3.0 (ALL)
2.38Average Toxicity (AvgTox)
3
MM-SafetyBench GPT-4o
2.8Average Toxicity
3