In the new release, request limits are per model. I tested Cloud4 for only 2 hours — without Max mode — and already hit my limit, without any visual alert.
Issue:
No per-model feedback or indicator – I can’t see how many requests I’ve used or how many remain.
Proposed quick improvements:
Progress bar per model , e.g.Cloud4: 120/150 requests used
Warning alerts when approaching limit, e.g.Only 10 requests left
Daily usage history panel with automatic reset at cycle start
Real-time update of usage upon stopping, so credits are reactivated immediately
Benefits:
Clear and predictable usage control
Prevents unexpected limit hits
Builds user confidence through transparency
A simple “health bar” for each model in the dashboard would really enhance usability. Thanks for considering!
@Michel_Zandonai limits are generic and not per model, limits are achieved faster using computational expensive models like sonnet4, o3, gemini2.5, when a limit occur, if its not a burst limit, it will free after a bunch of hours: Bad Usage Reporting on 3rd Party Extension - #24 by Zackh1998
+1 for an alert that limit will be reach soon: it allows to ‘close’ a session and prepare for the next
I saw a post from the team and in the official documentation that the system now suggests switching models once one hits its limit. Here’s what I experienced in practice:
I used Cloud4 with reasoning, and after just 2 hours, I hit the limit (without using Max mode).
Then I switched to Gemini 2.5 Pro (a premium model) and used it for about 3–4 hours with no limit.
Once I returned to Cloud4, it worked again.
This confirms that limits are applied per model—one model can hit its limit while another remains fully available, and the original model only unlocks again after usage stops.
