Anthropic has unveiled Claude Sonnet 4.6, describing it as the most powerful Sonnet model it has released so far.
The company says this upgrade is built to function as a dependable, everyday AI tool for developers and professionals.
According to Anthropic, Sonnet 4.6 shows clear improvements in coding accuracy, consistency, and instruction-following. Developers who received early access reportedly preferred it over the previous Sonnet version, and some even ranked it ahead of Claude Opus 4.5, which was considered its most advanced system late last year.
The improvements go well beyond coding. Anthropic says the model has been strengthened across computer use, long-context reasoning, agent planning, design tasks, and general knowledge work — areas that directly influence real-world productivity.
Like the recently launched Claude Opus 4.6, Sonnet 4.6 includes a 1 million token context window in beta. This expanded capacity allows users to work on longer and more complex sessions without resetting, making it more practical for extended projects.
The company also highlighted stronger desktop interaction skills, noting that the model performs more reliably when navigating and executing computer-based tasks. That progress brings it closer to handling structured office work that was previously associated with higher-tier models.
Pricing remains unchanged at $3 per million input tokens and $15 per million output tokens. Sonnet 4.6 is now the default model for Free and Pro users, while Opus continues as the premium offering.
Anthropic added that extensive safety testing found no major concerns, describing the model as warm, honest, and strongly aligned with responsible AI principles.
Also Read: ServiceNow Partners With Anthropic to Roll Out Claude to 29,000 Employees
