Claude Opus 4.1 Release Notes

Anthropic has released Claude Opus 4.1, an upgrade to Claude Opus 4 designed to improve performance in agentic tasks, real-world coding, and reasoning. This update serves as an incremental improvement ahead of larger model updates planned for the coming weeks.

Coding Performance and Benchmarks

Claude Opus 4.1 achieves a state-of-the-art coding performance score of 74.5% on SWE-bench Verified. The model demonstrates significant improvements in real-world software engineering tasks compared to its predecessor, Opus 4.

External partners have reported the following performance gains:

• GitHub: Notes improvements across most capabilities, with specific gains in multi-file code refactoring.
• Rakuten Group: Reports that Opus 4.1 excels at pinpointing exact corrections within large codebases without introducing bugs or making unnecessary adjustments.
• Windsurf: Reports a one standard deviation improvement over Opus 4 on their junior developer benchmark, comparing the performance leap to the transition from Sonnet 3.7 to Sonnet 4.

Research and Data Analysis Capabilities

Opus 4.1 enhances Claude's capabilities in in-depth research and data analysis. These improvements are specifically focused on detail tracking and agentic search, allowing for more precise information retrieval and synthesis.

Availability and Integration

Claude Opus 4.1 is available to paid Claude users and is integrated into Claude Code. For developers, the model is accessible via the API using the identifier claude-opus-4-1-20250805, as well as through Amazon Bedrock and Google Cloud's Vertex AI. Pricing for Opus 4.1 remains identical to the pricing for Opus 4.