AI Safety Index
67.86 / 100
3rd 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 | 91.97 / 100 | |
| Spiral-BenchMental & Emotional Safety | 69.11 / 100 | |
| KORAYouth Safety | 78.29 / 100 | |
| PatientSafetyBenchMedical Advice Safety | 96.46 / 100 | |
| HealthBench-HardMedical Advice Safety | 20.99 / 100 | |
| ELEPHANTManipulation | 28.12 / 100 | |
| DarkBenchManipulation | 51.66 / 100 | |
| HumanAgencyBench (Autonomy)Manipulation | 52.25 / 100 | |
| ASK — AI Scam KnowledgeSecurity | 57.85 / 100 | |
| FairMT-BenchBias & Fairness | 89.59 / 100 | |
| ConfAIdePrivacy / Confidentiality | 98.37 / 100 | |
| SimpleQA VerifiedMisinformation | 63.62 / 100 | |
| SYCON-BenchMisinformation | 84.40 / 100 | |
| HumanAgencyBench (Correct Misinformation)Misinformation | 85.28 / 100 | |
| SystemCheck / RealGuardrailsRule Following | 84.51 / 100 | |
| AgentIFRule Following | 57.06 / 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-sol · configuration ID openai.gpt-6-sol.medium-1ba09042bf0b