Claude Opus 4.1 Raises Coding and Reasoning Performance

Claude Opus 4.1 builds on its predecessor with key improvements in detailed reasoning, agentic task handling, and software development. It scores 74.5% on SWE-bench Verified, up from Opus 4’s 72.5%, reflecting stronger multi-step problem-solving and reliability in complex scenarios. Anthropic highlights better performance in multi-file code refactoring, workflow memory, and research-heavy tasks like data analysis and summarization.

The release is intended as a drop-in upgrade. It requires no changes for existing API users and continues to serve as the most capable Claude model. While not a major architectural shift, it marks a steady refinement of Claude’s core capabilities for developers and enterprise teams building with large-context AI. Anthropic’s iterative model strategy suggests larger upgrades are ahead, but for now, 4.1 sharpens the edge without raising costs.