AI Safety Index
65.37 / 100
8th 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 | 94.32 / 100 | |
| Spiral-BenchMental & Emotional Safety | 71.66 / 100 | |
| KORAYouth Safety | 78.29 / 100 | |
| PatientSafetyBenchMedical Advice Safety | 94.69 / 100 | |
| HealthBench-HardMedical Advice Safety | 31.16 / 100 | |
| ELEPHANTManipulation | 33.57 / 100 | |
| DarkBenchManipulation | 49.24 / 100 | |
| HumanAgencyBench (Autonomy)Manipulation | 31.09 / 100 | |
| ASK — AI Scam KnowledgeSecurity | 54.76 / 100 | |
| FairMT-BenchBias & Fairness | 75.95 / 100 | |
| ConfAIdePrivacy / Confidentiality | 98.00 / 100 | |
| SimpleQA VerifiedMisinformation | 40.14 / 100 | |
| SYCON-BenchMisinformation | 89.20 / 100 | |
| HumanAgencyBench (Correct Misinformation)Misinformation | 62.50 / 100 | |
| SystemCheck / RealGuardrailsRule Following | 76.15 / 100 | |
| AgentIFRule Following | 57.99 / 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 gpt-5.6-luna · configuration ID openai.gpt-5.6-luna.medium-a2c3a1aeeaa4
