Anthropic Expands Claude Sonnet 4.6 With 1M-Token Context and Stronger Coding Performance

New updates have been reported about Anthropic.

Anthropic is deepening its product stack with the beta release of Claude Sonnet 4.6, a mid-size model upgrade that will now serve as the default option for both Free and Pro users and is designed to improve coding, instruction-following, and computer-use tasks. The company is doubling Sonnet’s maximum context window to 1 million tokens, positioning the model to handle entire codebases, complex legal contracts, and large research corpora in a single session, which materially enhances its competitiveness for enterprise-scale workflows.

The Sonnet 4.6 launch follows the Opus 4.6 update by two weeks and signals Anthropic’s continued four-month model refresh cadence, with an updated Haiku expected next and reinforcing a clear roadmap across its model tiers. Benchmark results show record scores for Sonnet in computer-use and software-engineering tasks, and a 60.4% result on ARC-AGI-2 places it ahead of most peer models in reasoning, though it still trails Anthropic’s own Opus 4.6 and select flagship models from rivals, underscoring both its relative strength and remaining headroom. For executives, the release indicates Anthropic is sharpening its mid-range offering for cost-effective, high-context enterprise use cases, likely driving higher platform engagement and creating more upsell potential into its premium models as organizations standardize on Claude for complex knowledge and development workloads.