The model is now the default option for users on the Free and Pro plans. In beta, the LLM received a 1 million-token context window—double the previous limit—which Anthropic says can accommodate entire codebases, lengthy contracts, and dozens of research papers in a single prompt.

The release is accompanied by new benchmark records, including OSWorld (computer use) and SWE-Bench (software engineering tasks). Sonnet 4.6 also scored 60.4% on ARC-AGI-2, an abstract-reasoning benchmark, outperforming most rivals and trailing only Opus 4.6, Gemini 3 Deep Think, and one fine-tuned version of GPT-5.2.

Earlier in February, Anthropic upgraded its flagship Claude Opus to version 4.6, improving long-horizon planning, sustained task execution, and performance on large codebases. The company says capabilities that previously required an Opus-class model are now available in Sonnet 4.6.

Anthropic later reported raising $30 billion at a $380 billion valuation, with proceeds allocated to advanced research, product development, and infrastructure expansion. The company also said it conducted a large-scale safety evaluation of Sonnet 4.6, noting the new model is significantly more resistant to prompt-injection attacks than its predecessor.