AI Safety Index
66.06 / 100
6th of 14 published models
16 of 16 benchmarks · 100.0% of configured weight
Benchmark profile
Out of 100; higher is safer.
| Benchmark | Score visualization | Score out of 100 |
|---|---|---|
| SIM-VAILMental & Emotional Safety | 99.50 / 100 | |
| Spiral-BenchMental & Emotional Safety | 69.62 / 100 | |
| KORAYouth Safety | 70.14 / 100 | |
| PatientSafetyBenchMedical Advice Safety | 82.82 / 100 | |
| HealthBench-HardMedical Advice Safety | 38.57 / 100 | |
| ELEPHANTManipulation | 27.64 / 100 | |
| DarkBenchManipulation | 41.66 / 100 | |
| HumanAgencyBench (Autonomy)Manipulation | 44.62 / 100 | |
| ASK — AI Scam KnowledgeSecurity | 57.61 / 100 | |
| FairMT-BenchBias & Fairness | 74.14 / 100 | |
| ConfAIdePrivacy / Confidentiality | 98.62 / 100 | |
| SimpleQA VerifiedMisinformation | 73.62 / 100 | |
| SYCON-BenchMisinformation | 87.60 / 100 | |
| HumanAgencyBench (Correct Misinformation)Misinformation | 95.80 / 100 | |
| SystemCheck / RealGuardrailsRule Following | 82.00 / 100 | |
| AgentIFRule Following | 63.45 / 100 |
These are evaluations of API model configurations. They do not establish how a consumer app behaves with its own prompts, tools, or safeguards.
Exact configuration: model claude-opus-5-5 · configuration ID anthropic.claude-opus-5-5.medium-b9c82f60ef2d
