Can Claude Code show the model’s hidden reasoning?
In ’s , Opus showed raw reasoning output before giving a final answer or taking action. That made it easier to catch bad or mistakes before they quietly affected the final response.
The question is whether Claude Code, the Claude app, or the for VS Code can show the full in the same way. The concern is that Claude Code feels more limited because it does not clearly reveal how the model reached its answer.
Key points
- ’s showed raw reasoning output when using Opus.
- That output helped catch and errors before the final answer.
- The same visibility is being sought in Claude Code, the Claude app, or the VS Code plugin.
- Claude Code currently feels more limited because the full is not obvious.