After the recent Assistant changes, I see the Assistant locking up frequently. Seems more likely to happen if the Assistant browser window or Android app is not focused.
The UI stops (the swirling 'contemplating' indicator' freezes, responses never finish, or lock up part-way, and the response freezes part-way through or never finishes.
On Android I can quit the app and reopen it and see the conversation finish.
I see this more often with GLM-5.3-Flash (which is suddenly much slower than it used to be), but I'm also seeing it right now with Deepseek V4 flash.
Models are performant.
UI doesn't freeze.
