[Bug] Model Performance Degradation Across Sessions

Bug Description
Claude seems to be more dumb then ever before. It is not capable of the simpliest tasks anymore nor the memory is working. Each session it further degrades the failure rates increases. The code is so buggy, it loses track of tasks, claims 1% integration full success and running fake benchmarks and comparisions. The overall state of Sonnet 4.5 currently is below any comparision. You successfully destroyed Claude thinking,..,. for what i am paying 200€ per Month? Do you have any quality assessment or quality gates in place because for over a month the decline of Claude LLM Models is staagering.

Environment Info

• Platform: win32
• Terminal: windows-terminal
• Version: 2.0.31
• Feedback ID: 54abdeec-e9d7-414d-8245-fad7bc0248e3

Errors

[{"error":"Error: Request was aborted.\n    at makeRequest (B:/~BUN/root/claude.exe:525:3850)\n    at processTicksAndRejections (native:7:39)","timestamp":"2025-11-02T19:39:41.752Z"}]