AI Safety Index
70.15 / 100
1st 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 | 96.29 / 100 | |
| Spiral-BenchMental & Emotional Safety | 75.59 / 100 | |
| KORAYouth Safety | 81.81 / 100 | |
| PatientSafetyBenchMedical Advice Safety | 95.95 / 100 | |
| HealthBench-HardMedical Advice Safety | 33.41 / 100 | |
| ELEPHANTManipulation | 28.56 / 100 | |
| DarkBenchManipulation | 50.45 / 100 | |
| HumanAgencyBench (Autonomy)Manipulation | 54.99 / 100 | |
| ASK — AI Scam KnowledgeSecurity | 52.38 / 100 | |
| FairMT-BenchBias & Fairness | 85.45 / 100 | |
| ConfAIdePrivacy / Confidentiality | 98.22 / 100 | |
| SimpleQA VerifiedMisinformation | 75.37 / 100 | |
| SYCON-BenchMisinformation | 89.40 / 100 | |
| HumanAgencyBench (Correct Misinformation)Misinformation | 89.66 / 100 | |
| SystemCheck / RealGuardrailsRule Following | 79.91 / 100 | |
| AgentIFRule Following | 60.22 / 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-6-astra · configuration ID openai.gpt-6-astra.medium-72c320c75539
