AI Safety Index
60.38 / 100
12th 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 | 94.81 / 100 | |
| Spiral-BenchMental & Emotional Safety | 58.74 / 100 | |
| KORAYouth Safety | 43.82 / 100 | |
| PatientSafetyBenchMedical Advice Safety | 94.44 / 100 | |
| HealthBench-HardMedical Advice Safety | 27.71 / 100 | |
| ELEPHANTManipulation | 47.26 / 100 | |
| DarkBenchManipulation | 39.54 / 100 | |
| HumanAgencyBench (Autonomy)Manipulation | 44.00 / 100 | |
| ASK — AI Scam KnowledgeSecurity | 67.61 / 100 | |
| FairMT-BenchBias & Fairness | 51.91 / 100 | |
| ConfAIdePrivacy / Confidentiality | 97.62 / 100 | |
| SimpleQA VerifiedMisinformation | 69.44 / 100 | |
| SYCON-BenchMisinformation | 99.00 / 100 | |
| HumanAgencyBench (Correct Misinformation)Misinformation | 43.44 / 100 | |
| SystemCheck / RealGuardrailsRule Following | 93.72 / 100 | |
| AgentIFRule Following | 62.44 / 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 gemini-3.6-flash · configuration ID google_cloud.gemini-3.6-flash.medium-fea0ca9a66f0
