Claude Sonnet 4.6 is here: Coding giant with million-token context window

Anthropic, the AI giant behind the popular Claude AI chatbot, has launched Sonnet 4.6, the latest version of its mid-sized Sonnet model, keeping up with its commitment to a four-month update cycle.

Part of Anthropic's Sonnet 4.6 announcement was a highlight on significant enhancements in coding capabilities, instruction-following, and overall computer use.

Key features of Anthropic's new Sonnet 4.6

The headlining detail of the new Sonnet version is that Sonnet 4.6 will serve as the default model for users on both the Free and Pro plans, with one of the most notable features being its beta version comprising an impressive context window of 1 million tokens—double compared to the previous largest window available for Sonnet.

This new context window, as described by Anthropic, is sufficient to accommodate entire codebases, lengthy contracts, or multiple research papers in a single request.

The new Sonnet variant comes on the heels of Opus 4.6, which was released just two weeks ago, and an updated Haiku model is also on the horizon, likely to be announced in the coming weeks.

Sonnet 4.6 also sets new benchmark scores, achieving high marks in OS World for computer use and SWE-Bench for software engineering. What reflects its impeccable capabilities is its 60.4% score on the ARC-AGI-2 benchmark, which evaluates skills related to human intelligence.

While this score puts Sonnet 4.6 ahead of many competitive models, it still lags behind Opus 4.6, Gemini 3 Deep Think, and an enhanced version of GPT 5.2.