The company Anthropic presented the Claude Sonnet 4.5 model, Anthropic reported. The new version is being evaluated as the most powerful model for writing code and completing complex tasks on a computer. The model has also achieved great success in logical thinking, mathematics, and domain-specific knowledge.
Together with Claude Sonnet 4.5, a range of new features were also introduced: a checkpoint function in the Claude Code application, an updated terminal interface, and a dedicated extension for VS Code. In addition, through the Claude API, the opportunity was created for agents to perform long-term memory and tasks of high complexity.
At the same time, the Claude Agent SDK was presented for developers. This platform is the very infrastructure powering Claude Code, and now allows the creation of AI agents in any field.
Read “Zamin” on Telegram!
