Join us as we explore the release of Anthropic’s Claude Sonnet 4.6, the new default model that brings Opus-level intelligence to a more accessible price point. In this episode, we break down the groundbreaking "computer use" capabilities, where the model utilizes an agentic loop to navigate complex spreadsheets and web forms with near-human performance.

We also analyze new benchmark data showing how Sonnet 4.6 outperforms GPT 5.2 in code review balance and why it is notably less prone to overengineering than its predecessors. Finally, we discuss the implications of the 1 million token context window and how enterprise teams are using these new tools to automate entire codebases and financial workflows at scale