Anthropic launched Claude Sonnet 5 on June 30, 2026, positioning it as its most agentic Sonnet model for coding, browser use, planning, and knowledge work. Anthropic’s release notes list a 1 million-token context window, 128,000-token maximum output, and promotional API pricing through August 31, while the company’s...
Anthropic introduced Claude Sonnet 5 on June 30, 2026, describing the model as its most agentic Sonnet release to date.
In its announcement, Anthropic said Claude Sonnet 5 is designed to improve reasoning, tool use, coding, and knowledge work. The company’s platform release notes list the model ID as claude-sonnet-5 and describe the release as available from June 30, 2026.
Anthropic’s documentation says the model supports a 1 million-token context window and a maximum output length of 128,000 tokens. Those limits position Sonnet 5 for long-document analysis, multi-step coding tasks, and workflows where a model needs to retain substantial project context.
Axios also reported that Anthropic released Sonnet 5 as a lower-priced model for agentic AI tasks, including browser use, planning, coding, and knowledge work. According to Axios, Anthropic is making Sonnet 5 the default model for Claude Free and Pro users.
Anthropic’s Claude Platform release notes list API pricing for Claude Sonnet 5 at $2 per million input tokens and $10 per million output tokens through August 31, 2026. The same release notes say pricing then changes to $3 per million input tokens and $15 per million output tokens.
The release notes also describe migration behavior changes for developers moving to the new model. Anthropic’s documentation is the primary source for those implementation details, including the model identifier, context length, output limits, and pricing schedule.
For customers building applications on Anthropic’s API, those details matter because model naming, context size, output caps, and token pricing affect cost estimates and integration work. Anthropic’s own documentation should be treated as the reference for current availability and configuration.
Anthropic published a Claude Sonnet 5 System Card alongside the release. The document includes benchmark results across software engineering, terminal tasks, browsing, general reasoning, and computer-use evaluations.
The system card lists evaluations including SWE-bench Pro, Terminal-Bench 2.1, BrowseComp, Humanity’s Last Exam, and OSWorld-Verified. Anthropic presents these results as support for its claims that Sonnet 5 improves on agentic and knowledge-work capabilities.
As with any model benchmark, the system card’s results should be read in context. Benchmarks can help compare capabilities under controlled conditions, but they do not fully predict real-world reliability, latency, cost, or task success in every production environment. Anthropic’s system card is still an important source because it documents the company’s own testing and the evaluation areas it considers relevant for Sonnet 5.
Anthropic is positioning Sonnet 5 as a practical model for broad use rather than only as a frontier-tier option for specialized cases. Axios characterized the release as focused on everyday work, and Anthropic’s announcement emphasizes coding, browser use, planning, and knowledge tasks.
For individual Claude users, the most visible change may be default access in Claude Free and Pro, as reported by Axios. For developers, the most concrete changes are in Anthropic’s platform documentation: the claude-sonnet-5 model ID, 1 million-token context, 128,000-token maximum output, and the stated API pricing schedule.
The release expands Anthropic’s Sonnet line with a model the company says is more capable at tool use and multi-step work. The strongest source-backed takeaway is narrower than the marketing language: Anthropic has launched Claude Sonnet 5, documented its API characteristics, published benchmark support in a system card, and is making the model available as a lower-priced option for common agentic AI tasks.
Anthropic introduced Claude Sonnet 5 on June 30, 2026, describing the model as its most agentic Sonnet release to date.
A new Sonnet model aimed at everyday agentic work In its announcement, Anthropic said Claude Sonnet 5 is designed to improve reasoning, tool use, coding, and knowledge work.
The company’s platform release notes list the model ID as claude sonnet 5 and describe the release as available from June 30, 2026.
Continue reading
