The pro weekly limit was increased nine times before, but it seems to have been changed back again these past couple of days. Not sure if it’s a bug or intentional. Hopefully it’s just a bug and not another silent change.
I ran into the quota limit/weekly quota/blocked from the service myself. What I did to determine the source of the problem was to use other products in sequence, in my daily workflow, to see if I would hit quota limits as fast, or hit weekly caps, etc. I used two other LLM products (inside Antigravity/Antigravity IDE when it got released). The answer: not by a long shot. The settings and options for reasoning, strength, and the model variants (low/medium/high) are consistent enough to compare them against each other, and Gemini is the only product that gives me this quota consumption headache. I never got blocked from using other products due to exhausting the 5-hour/weekly quota limit; all that happens is I have to wait for the cooldown period to end, and resume work. I have been doing it for a few months now, and I am highly confident that the problem is with Google’s product, not the way I use it.
Im getting hit with a 4 days cooldown even though previous UI saying it’s X Hours until refresh. Is this a new form of scamming?
Yeah, the UI is pretty awful. My guess is there is some kind of job that runs, sets the hourly countdown but at some point this gets overriden by some other job that checks against the weekly quota and the UI goes from saying x hours to y days. Really annoying.
sorry for the late reply but this is not actually the issue.
the issue itself at its core is:
how can I investigate where is the issue if the only thing available for users is the % of the token usage?
They updated the usage percentage to reflect the weekly quota, which is much better*. However, I don’t know if there will be a future update that allows me to view token usage for each message, so I can investigate where the problem occurred and this is what I need since it will be good for us developers and for Google as well so we can report and provide feedback.
*this is way better:
