Anthropic Launches Claude Sonnet 4.6 With Double the Context Window and Record-Breaking Benchmark Scores
Summary
Anthropic launches Claude Sonnet 4.6, doubling its context window to 1 million tokens and setting new benchmark records with a 60.4% score on ARC-AGI-2, making it the new default model for Free and Pro users.
Key Points
- Anthropic releases Sonnet 4.6, the latest update to its midsized model, featuring improvements in coding, instruction-following, and computer use, and will serve as the default model for Free and Pro plan users.
- Sonnet 4.6 debuts with a 1 million token context window — double its predecessor — capable of holding entire codebases, lengthy contracts, or dozens of research papers in a single request.
- The model sets new benchmark records, including a 60.4% score on ARC-AGI-2, placing it above most comparable models, though it still trails Opus 4.6, Gemini 3 Deep Think, and a refined version of GPT 5.2.