anthropic shipped claude sonnet 4.6 on february 17 as a full sonnet-line upgrade for coding, long-context reasoning, computer use, and office work. on the api side, the release also brought a 1m-token context window in beta.
that matters because sonnet is usually the practical everyday claude choice, not the prestige tier. if you build tools, debug larger codebases, or use claude for structured production work, this is the kind of model update that changes default behavior rather than just adding a new checkbox.
anthropic also kept pricing aligned with the previous sonnet tier. that makes the release more usable for sustained workflow automation, not just one-off benchmarks or demo prompts.