Anthropic Rolls Out Claude Sonnet 4.6 as Default Model, Boosts Coding and Long-Context Reasoning

Anthropic has introduced Claude Sonnet 4.6, describing it as its most advanced Sonnet model to date, with significant gains in coding, reasoning, and handling extensive datasets. The new model is now the default option for users of the Claude chatbot, available to both free and Pro subscribers, marking another rapid step forward in the intensifying global AI race.

The launch follows closely on the heels of Claude Opus 4.6, underscoring Anthropic’s fast-paced development strategy. With Sonnet 4.6, the company is aiming to refine everyday productivity tasks while also strengthening high-level technical capabilities.

Currently, Sonnet 4.6 is accessible directly within Anthropic’s chatbot platform. Free-tier users can interact with the model under usage limits that reset every five hours, while Pro subscribers retain access under existing subscription pricing. Beyond individual users, Anthropic has also made the model available via API and on major cloud platforms, allowing developers and enterprises to integrate it into their own applications and workflows.

“It’s a full upgrade of the model’s skills across coding, computer use, long-context reasoning, agent planning, knowledge work, and design. Sonnet 4.6 also features a 1M token context window in beta,” says the company in its official announcement blog post.

According to Anthropic, one of the most noticeable improvements lies in coding performance. The company says the new model follows programming instructions more reliably and produces cleaner, more usable code without deviating from the prompt—an issue that has challenged many AI systems in the past. It is also said to perform better at editing and debugging code, offering greater consistency across technical tasks. Early feedback from testers suggests a preference for Sonnet 4.6 over earlier versions, and in some internal agent-based evaluations, the model reportedly outperformed Claude Opus 4.6.

Beyond software development, Sonnet 4.6 is positioned as a stronger tool for “knowledge work.” This includes analysing lengthy documents, summarising reports, working with spreadsheets, and assisting in design-related workflows. A major contributor to this capability is the model’s 1 million token context window, currently in beta. With this expanded memory capacity, the AI can process and retain far more information within a single session—an advantage for reviewing legal contracts, financial documents, research material, or large codebases.

On performance benchmarks, Anthropic reports strong results across industry-standard evaluations such as Humanity’s Last Exam, GPQA Diamond, and SWE-bench Verified, which assess reasoning depth and coding accuracy. The company also highlights improved outcomes in enterprise automation and insurance-related tasks compared to earlier Claude releases.

Safety remains part of the upgrade. Anthropic states that Sonnet 4.6 demonstrates lower hallucination rates and reduced “sycophancy,” meaning it is less likely to agree with incorrect user assumptions.

The release comes as competition intensifies among AI leaders. Companies like OpenAI and Google continue to accelerate updates to their flagship systems, pushing advancements in reasoning, coding, and multimodal performance. With new model iterations emerging in rapid succession, Anthropic’s latest release signals that the race to build more capable and reliable AI systems is only gaining speed.