65K TPM is not enough, I end up hitting limits on it for my use cases where its Audio + video/screenshots and using up that quota, the older models have 1M TPM whereas the new one doesn’t? I feel like we should at least get 128k so its more usable and doesn’t get quota limited, i get the model just released but maybe drop down the quota on the old model for the free tier and give some love to the new model so developers can start moving to it instead
Found a solution, i was sending the screenshots at HIGH resolution which is 1120 tokens per image, which is A LOT after ~60s, Started sending them at LOW and that made a good improvement