Anthropic has officially rolled out Claude Opus 4.1, the latest upgrade to its flagship AI model, now accessible to paying users through Claude Code, API, Amazon Bedrock, and Google Cloud’s Vertex AI—with no pricing changes from previous versions.
Building on the strengths of Opus 4, the 4.1 release brings notable enhancements in software development, reasoning, and agentic task execution. The most standout improvement is the model’s state-of-the-art coding capabilities, where it has achieved a 74.5% score on SWE-bench Verified, a highly regarded benchmark that evaluates how well models perform on real-world software engineering issues.
According to GitHub, Claude Opus 4.1 has made measurable gains across most coding-related tasks, showing particular strength in multi-file code refactoring—a complex challenge for AI systems. This places Opus 4.1 among the top performers in the domain of AI-assisted software engineering.
Organizations testing the model are already seeing the impact. Rakuten Group shared that the model demonstrated surgical accuracy when applied to large-scale codebases—correcting issues without introducing bugs or unnecessary edits. Similarly, AI startup Windsurf reported a one standard deviation improvement in task performance compared to Opus 4, particularly on its internal junior developer benchmark, further reinforcing its reliability in production-level environments.
To upgrade, developers can access the new version using the identifier: claude-opus-4-1-20250805.
Anthropic noted that this release is just the beginning of a broader roadmap. “We plan to release substantially larger improvements to our models in the coming weeks,” the company said, encouraging users to explore the detailed documentation, system cards, and model pages now available. Feedback from the developer community is also being actively sought to shape future iterations.
As competition in the AI model space intensifies, Claude Opus 4.1 positions itself as a powerful, developer-centric solution for advanced coding and real-world task orchestration—without added cost.