Statement from Coin News: Anthropic launched the Claude Sonnet 5 model, stating that it is the most powerful Sonnet model in terms of agentic capabilities to date, as it can make plans and use tools such as a browser and terminal, and operate independently at a level that was previously only achieved by larger and more expensive models a few months ago. Sonnet 5's performance approaches that of Opus 4.8, while its price is lower, and it has significantly improved over the previous generation Sonnet 4.6 in key agent dimensions such as reasoning, tool use, programming, and cognitive work. Safety evaluation showed that the rate of harmful behaviors in Sonnet 5 is lower than in Sonnet 4.6, and it is safer in agent scenarios, and its capabilities in cybersecurity tasks are much lower than the current Opus model. Starting today, Sonnet can be used.