Anthropic have announced the release of Claude Opus 4.1, an upgraded version of Claude Opus 4 that delivers improved performance on agentic tasks, real-world coding, and reasoning capabilities.
The new model achieves 74.5% performance on SWE-bench Verified, a popular coding evaluation benchmark. Claude Opus 4.1 is immediately available to paid Claude users, through Claude Code, and via Anthropic's API, Amazon Bedrock, and Google Cloud's Vertex AI platform. Pricing remains unchanged from Claude Opus 4.
GitHub reported that Claude Opus 4.1 demonstrates improvements across most capabilities compared to its predecessor, with particularly notable gains in multi-file code refactoring. Enterprise customer Rakuten Group highlighted the model's precision in identifying exact corrections within large codebases while avoiding unnecessary adjustments or bug introduction. The company's team expressed preference for this precision in everyday debugging tasks.
Development platform Windsurf documented a one standard deviation improvement over Opus 4 on their junior developer benchmark, describing performance gains comparable to the leap from Sonnet 3.7 to Sonnet 4. The model also enhances in-depth research and data analysis capabilities, particularly in detail tracking and agentic search functionality.
Anthropic indicated plans to release substantially larger model improvements in coming weeks. The company recommends all users upgrade from Opus 4 to the new version. Developers can access the model through the API using the identifier claude-opus-4-1-20250805.
Organisations utilising Claude for coding tasks can expect enhanced multi-file refactoring capabilities and improved debugging precision. The model's enhanced research and data analysis features support enterprise-scale analytical workflows requiring detailed information tracking.
Claude Opus 4.1 positions enterprises for improved development productivity through enhanced coding accuracy and reduced debugging time. The maintained pricing structure allows organisations to upgrade without budget adjustments while accessing improved technical capabilities that support complex software development initiatives.