Anthropic, an AI research company, has developed a safer large language model (LLM) called Claude, which surpasses other popular models like ChatGPT and GPT-4o on various benchmarks. Claude 3.5 Sonnet is the latest version of Anthropic's LLM, offering improved capabilities such as coding prowess, vision capabilities, and a lower price point compared to its predecessor. The company's approach to AI safety includes Constitutional AI, red teaming, and being a public benefit corporation, which empowers them to prioritize safety over profits. Claude is designed to work well on open-ended questions, provide helpful advice, and perform tasks like writing, editing, and summarizing text. Anthropic aims to influence the development of safer AI systems by competing commercially and raising the bar for safety standards.