OpenAI has launched the Safety Evaluations Hub, a dedicated webpage that publicly displays the safety performance of its artificial intelligence models. The hub provides detailed results on how these models fare in tests for harmful content, hallucinations, and jailbreak vulnerabilities. This initiative aims to enhance transparency and accountability in AI development. In addition, OpenAI introduced HealthBench, a benchmark designed to assess the safety and accuracy of large language models like ChatGPT in handling healthcare-related queries. HealthBench uses physician-designed real-world scenarios to evaluate AI's ability to suggest treatments and support medical professionals. These developments reflect OpenAI's ongoing efforts to improve the reliability and safety of AI applications, particularly in sensitive areas such as healthcare. The Safety Evaluations Hub and HealthBench represent steps toward more rigorous and publicly accessible evaluation standards for AI models.