Anthropic has announced the release of its most advanced AI models to date – Claude Opus 4 and Claude Sonnet 4 – which focus on complex and sustained tasks.
Whilst much of the discussion in the industry remains centered on ChatGPT and Microsoft’s Copilot, both of which are powered by OpenAI models, Anthropic has strengthened Claude’s ability to perform long-running processes such as coding, research, data analysis, and technical writing.
Claude Opus 4 is positioned as the company’s most capable model, designed to handle extended workloads and complex instructions, including writing and maintaining large codebases. Sonnet 4, a smaller but upgraded version of its predecessor, also has improvements in reasoning, coding, and instruction-following.
Both models include support for tool use and can work with local files to extract and retain key information across sessions. This aims to enable more agent-like behaviour, where the model can be used to support or automate parts of larger workflows rather than just answer individualised queries. Anthropic has also added new developer features to its API, such as code execution tools, prompt caching, and file handling capabilities – all features aimed at enterprise automation and developer use cases.
This release is part of Anthropic’s broader push to establish its models as reliable, high-performance agents rather than purely conversational tools. Smaller and newer than rival OpenAI, it has also tried to position itself as the safest AI model. However, the company has not escaped controversy – just last week, Reddit filed a lawsuit in San Francisco alleging that Anthropic used Reddit user content without consent to train its models. The company also released a report accompanying the launch which acknowledges the potential of Claude Opus 4 to resort to ‘extremely harmful actions.’