1
Verify the likely cause
Reduce context length, summarize older turns, use a smaller quantization and monitor VRAM allocation while reproducing the slowdown.
FULL AI ANALYSIS
Local LLM slows down dramatically after a long conversation
Generated Sep 26, 2026, 11:55:04 AM
Analysis summary
The reported behaviour is consistent with the key-value cache outgrows available VRAM and layers spill to system memory. The safest next step is to verify that hypothesis with a reversible test before making permanent configuration changes.
Possible causes
The key-value cache outgrows available VRAM and layers spill to system memory.
87%A recent software, firmware or configuration change is exposing an existing compatibility issue.
39%Suggested fixes
2
Validate and monitor
Reproduce the issue once after the change, review relevant logs or metrics, and keep a rollback path until the system has remained stable.