CtrlK
Claude Opus 4.7 is a hybrid reasoning model from Anthropic announced on April 16, 2026. Anthropic describes it as its most capable generally available model and positions it for coding, multi-step agent workflows, vision tasks, and enterprise knowledge work.[1]
The public product page describes Claude Opus 4.7 as a 1 million-context model that improves on Opus 4.6 in coding, vision, and difficult multi-step tasks. Anthropic says the model is more thorough and more consistent on hard work than earlier Opus releases.[1]
| Area | Details |
|---|---|
| Announcement date | April 16, 2026 |
| Context window | 1 million tokens |
| Positioning | Anthropic's most capable generally available model |
| Main use cases | Software engineering, agent workflows, enterprise documents, and professional knowledge work |
| Deployment surfaces | Claude, Claude API, Amazon Bedrock, Google Cloud Vertex AI, Microsoft Foundry |
Anthropic also says Opus 4.7 uses adaptive thinking, which means the model adjusts how much reasoning it uses based on task difficulty rather than forcing a fixed thinking budget for every request.[1][2]
Opus 4.7 launched for Claude Pro, Max, Team, and Enterprise users, and on the Claude Platform for developers.[1]
Anthropic listed the API price at $5 per million input tokens and $25 per million output tokens, with prompt caching discounts up to 90% and a 50% batch-processing discount. The company also offered a US-only inference option at 1.1x normal pricing for customers that need US residency.[1]
The Opus 4.7 page places the model above Opus 4.6, Opus 4.5, Opus 4.1, and the original Opus 4 release. Anthropic's examples and customer statements focus on long-running coding jobs, document-heavy analysis, and sustained tool use across multi-step workflows.[1]
Claude Opus 4.7 is part of Anthropic's larger move from chatbot-style interaction toward persistent agentic work. The official description emphasizes that the model can plan carefully, run for longer, learn across sessions with memory features, and handle large codebases and business documents with less supervision.[1][2]