Jun 12
Fixing Screen Freezing Issues in Large Context Environments
When interacting with LLMs (particularly Thinking models), after approximately 4–5 rounds of follow-up exchanges, submitting a prompt causes the rendering speed to drop significantly — specifically during model output — and the screen appears to freeze. This issue tends to occur more readily in situations where the context has grown large, such as after repeated deep research sessions.Starting a new chat resolves the issue, but similar problems rarely occur with other services.
Completed