[Bug] Claude model exhibits persistent reasoning bias and fails to accept contradicting information

Question

**Bug Description**
Opus 4.6 is great, really good reasoning and no problem at all, expect for the fact that sometimes he creates some barriers in his minds, and even if you tell him something, he's stuck with that idea and you can't change it.

**Environment Info**
- Platform: win32
- Terminal: pycharm
- Version: 2.1.101
- Feedback ID: 972deda6-0fcb-423b-9fc3-337bc9b8962c

**Errors**
```json
[]
```