Anthropic Releases Claude Opus 4.7 With Smarter Coding, Adaptive Reasoning, and New Effort Controls

Anthropic releases Claude Opus 4.7, its most powerful coding and agentic model yet, featuring adaptive reasoning that dynamically allocates thinking resources, a new 'xhigh' effort tier, and smarter bug detection — but users must adjust prompting habits as the model defaults to leaner, less verbose responses than its predecessor.

• Claude Opus 4.7 is now the strongest generally available model for coding and agentic tasks, offering improved bug detection, better context retention across sessions, and enhanced reasoning through ambiguous problems compared to Opus 4.6.
• A new default effort level called 'xhigh' is being introduced between 'high' and 'max,' recommended for most agentic coding work, while adaptive thinking replaces fixed thinking budgets, allowing the model to dynamically allocate reasoning tokens based on task complexity.
• Users are advised to front-load task details in the first prompt, reduce back-and-forth interactions, and explicitly guide tool use and subagent spawning, as Opus 4.7 defaults to less verbose responses, fewer tool calls, and more conservative subagent delegation than its predecessor.