AI Briefing4.7, Agent Evolution
Key point
Claude Opus 4.7 significantly strengthens long-horizon tasks, self-verification, and vision performance.
Details
Anthropic has unveiled Claude Opus 4.7, available right now on claude.ai, the platform, and major clouds.
The core change is enhanced autonomy on long-running tasks. It handles longer workflows more rigorously, follows instructions more precisely, and self-verifies before delivering final results.
Vision performance has also improved significantly. It processes images at more than 3x higher resolution than the previous version, which is expected to bring noticeable improvements when analyzing visual materials such as interfaces, slides, documents, and charts.
On the API side, the following features have been added.
- xhigh effort level: an option to control reasoning depth/latency between
highandmax - task budgets (beta): supports cost management and priority adjustment during long-running sessions
Claude Code has also been updated.
- /ultrareview: a dedicated code review command at a level of thoroughness comparable to human review
- Auto mode has been expanded to Max users, enabling long tasks to be handled with fewer interruptions
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
US Appeals Court Upholds Pentagon Blacklisting of Anthropic as Supply Chain Risk