AI Safety Index
65.41 / 100
7th 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 | 96.79 / 100 | |
| Spiral-BenchMental & Emotional Safety | 65.77 / 100 | |
| KORAYouth Safety | 62.95 / 100 | |
| PatientSafetyBenchMedical Advice Safety | 98.98 / 100 | |
| HealthBench-HardMedical Advice Safety | 31.37 / 100 | |
| ELEPHANTManipulation | 38.25 / 100 | |
| DarkBenchManipulation | 50.90 / 100 | |
| HumanAgencyBench (Autonomy)Manipulation | 57.74 / 100 | |
| ASK — AI Scam KnowledgeSecurity | 51.19 / 100 | |
| FairMT-BenchBias & Fairness | 69.39 / 100 | |
| ConfAIdePrivacy / Confidentiality | 98.22 / 100 | |
| SimpleQA VerifiedMisinformation | 50.94 / 100 | |
| SYCON-BenchMisinformation | 99.40 / 100 | |
| HumanAgencyBench (Correct Misinformation)Misinformation | 67.90 / 100 | |
| SystemCheck / RealGuardrailsRule Following | 86.19 / 100 | |
| AgentIFRule Following | 57.68 / 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 grok-4.5 · configuration ID xai.grok-4.5.medium-edb165a2df11
