Anthropic has released Sonnet 4.6, a new version of its midsized Sonnet model. This release aligns with the company’s four-month update cycle.
Sonnet 4.6 will serve as the default model for users with Free and Pro plans.
Key improvements highlighted by Anthropic for Sonnet 4.6 include advancements in coding capabilities, instruction-following, and general computer use.
The beta version of Sonnet 4.6 features a 1 million token context window. This capacity is double the size of the previous largest window available for Sonnet models. Anthropic stated this window can accommodate “entire codebases, lengthy contracts, or dozens of research papers in a single request.”
This launch follows the release of Opus 4.6 two weeks prior. An updated Haiku model is anticipated in the near future.
Sonnet 4.6 achieved new record benchmark scores in several areas:
- OS World for computer use
- SWE-Bench for software engineering
The model also scored 60.4% on ARC-AGI-2, a benchmark designed to assess human intelligence-specific skills. This score places Sonnet 4.6 ahead of most comparable models. However, it trails models such as Opus 4.6, Gemini 3 Deep Think, and one refined version of GPT 5.2.
Featured image credit
