Anthropic releases Claude Opus 4.1 with coding, reasoning, and debugging improvements

Anthropic has unveiled Claude Opus 4.1, positioning it as a major upgrade focused on agentic tasks, real-world coding, and reasoning improvements. This update aims to deliver better results for users requiring advanced research and precise code manipulation.

Paid users now have access to Opus 4.1 in both Claude Code and through Anthropic’s API. Following its broader strategy, the model is also available via Amazon Bedrock and Google Cloud’s Vertex AI, all at the same pricing as Opus 4.

Claude Opus 4.1 advances technical capabilities, reaching 74.5% on the SWE-bench Verified benchmark for coding performance. Along with these gains, the model reportedly strengthens skills in in-depth research, detail tracking, and agentic search, expanding its utility for data analysis professionals.

Feedback from developer communities helps validate these upgrades. GitHub reports notable advances across most areas relative to Opus 4, especially for multi-file code refactoring tasks. Rakuten Group highlights that Opus 4.1 can pinpoint corrections within large codebases without introducing unnecessary changes or new bugs, supporting efficient debugging. Windsurf adds that Opus 4.1 delivers a one standard deviation improvement over Opus 4 on a junior developer benchmark, reflecting gains comparable to previous Sonnet model upgrades.