Anthropic has announced the general availability of its latest model, Claude Opus 4.7, marking a significant upgrade over its predecessor, Opus 4.6. The new model delivers notable improvements in advanced software engineering, particularly for complex and long-duration coding tasks that previously required close human oversight.

According to the company, Opus 4.7 demonstrates stronger reasoning capabilities, enhanced consistency, and improved adherence to user instructions. It is also capable of self-verification, designing internal checks to validate outputs before delivering results, making it more reliable for high-stakes development workflows.

Beyond coding, the model introduces substantial advancements in vision capabilities, allowing it to process higher-resolution images and generate more refined outputs across professional tasks such as user interfaces, presentations, and documentation. While it does not surpass the experimental Claude Mythos Preview in overall capability, Opus 4.7 outperforms Opus 4.6 across a wide range of benchmarks.

The release follows Project Glasswing, an initiative aimed at exploring both the risks and benefits of AI in cybersecurity. As part of its safety-first rollout strategy, Anthropic is introducing Opus 4.7 with built-in safeguards designed to detect and block potentially harmful or high-risk cybersecurity requests. These measures will be tested in real-world deployments before the broader release of more advanced models like Mythos.

To support legitimate security use cases, the company has also launched a Cyber Verification Program. This initiative enables vetted professionals to access the model for activities such as vulnerability research, penetration testing, and red-teaming.

Claude Opus 4.7 is now available across Anthropic’s ecosystem, including its API, as well as major cloud platforms such as Amazon Bedrock, Google Cloud Vertex AI, and Microsoft Foundry. Pricing remains unchanged from the previous version, set at $5 per million input tokens and $25 per million output tokens.

The launch signals Anthropic’s continued focus on balancing cutting-edge AI performance with robust safety frameworks, particularly as models become more capable in sensitive domains like cybersecurity.