Anthropic Unveils Claude Opus 4.1 with Enhanced Coding and Reasoning Capabilities

Anthropic has launched Claude Opus 4.1, a significant upgrade to its predecessor, Claude Opus 4, enhancing capabilities in agentic tasks, real-world coding, and reasoning. This new version is now accessible to paid Claude users and is integrated with Claude Code. Additionally, it is available through Anthropic's API, Amazon Bedrock, and Google Cloud's Vertex AI, with pricing remaining consistent with Opus 4, according to Anthropic.

Key Improvements in Claude Opus 4.1

Claude Opus 4.1 showcases a remarkable advancement in state-of-the-art coding performance, achieving a 74.5% success rate on the SWE-bench Verified platform. The model has also enhanced its research and data analysis capabilities, particularly in detail tracking and agentic search, making it a valuable tool for developers and researchers alike.

GitHub has noted improvements across most capabilities of Claude Opus 4.1 compared to its predecessor, with significant gains in multi-file code refactoring. Rakuten Group has highlighted the model's precision in identifying exact corrections within large codebases, avoiding unnecessary changes or bug introductions. This precision is particularly valued for everyday debugging tasks. Moreover, Windsurf reports a notable one standard deviation improvement over Opus 4 on their junior developer benchmark, akin to the performance leap from Sonnet 3.7 to Sonnet 4.

Availability and Usage

Anthropic recommends users upgrade from Opus 4 to Opus 4.1 to leverage its improved capabilities. Developers can access the new version using the code claude-opus-4-1-20250805 via the API. Further information on the system's capabilities, models, and pricing can be found on Anthropic's official website, offering insights for both current and prospective users.

The release of Claude Opus 4.1 marks a significant step in Anthropic's ongoing efforts to develop reliable, interpretable, and steerable AI systems. The company plans to introduce further enhancements to its models in the coming weeks, aiming to continuously improve AI performance and utility.