ResearchBenchmarksFeasible Certification on Drug formulationFollow0.9Certified Success RatePHACT0.8550.87750.90.9225Jun 29, 2026Evaluation ResultsMethodMethodLinksCertified Success RateLoop FailuresWrong VerdictsTimeoutsPHACTModel=Gemini 2.5 Flash...Model=Gemini 2.5 Flash, Temperature=02026.060.9010