Anthropic Releases Opus 4.7

Anthropic has launched Claude Opus 4.7, positioning it as a significant step forward for complex software engineering tasks. The model demonstrates particular strength on the most difficult coding challenges, with users reporting they can now hand off work that previously required close supervision. Beyond coding, Opus 4.7 offers substantially improved vision capabilities, processing images at greater resolution, and produces more tasteful, creative outputs for professional documents, interfaces, and presentations.

A key differentiator is Anthropic's approach to cybersecurity safeguards. While less capable than the unreleased Claude Mythos Preview model in cyber operations, Opus 4.7 serves as a testing ground for new safety measures. The company has implemented automatic detection and blocking of prohibited or high-risk cybersecurity requests. Security professionals seeking legitimate use for vulnerability research can apply through Anthropic's new Cyber Verification Program, with learnings from this deployment informing future model releases.

The model is now generally available across Anthropic's entire product ecosystem, including the Claude API, Amazon Bedrock, Google Cloud's Vertex AI, and Microsoft Foundry. Pricing remains unchanged from Opus 4.6 at $5 per million input tokens and $25 per million output tokens, making this enhanced capability accessible without cost increase. This release represents Anthropic's continued strategy of iterative improvement while implementing real-world testing of critical safety protocols.