when I run antigravity today, I saw announcement: “If you are on a paid plan, your Gemini quota has been reset for the week and increased by 3x moving forward. Keep building!”
thanks for it, now I m planning to move Ultra
but the quota increases is for gemini or calude or both ?
the quota actually did not increase, it’s just shared between the models, so if before, your FLASH model doesn’t run out of juice EVER then now the PRO can utilize the FLASH’s quota to resupply the PRO’s own quota. from what I’ve observed, it’s the same for the GPT and claude, before, ONE, 5 hour reset cannot refill the entire quota for GPT and claude if you exhaust all of it, but now it can fill it up to max:
but here’s the catch, if you notice one of them not refilling at all, that means you’re running out of quota for all models for the week, BUT models’ consumption rate are not equal, meaning if you have 100 quota (not 100%, but a compute value of 100, just a numerical representation example, we don’t actually know how google calculates compute for quota) 100 quota left, flash model is cheaper in compute consumption, so it will for example consume only 5 per message which PRO/claude will consume 25 per message for the same prompt. but it doesn’t change the fact the you are running out of compute and you should use it wisely, whether use flash for 20 minor changes or use Claude for 4 major prompts, either way, it’s shared not increased they never mentioned increasing it, just a change in computation. Apparently, there is no saving compute constraints, demand is too high to increased it (see quote) i also think this one is true, after all major AI providers suddenly decreasing usage. This time google really rebuilt their business model to an acceptable level…? but its still too early to say
so before, it might have been each model uses a different API which has separate quota for each API, now all models use the same API so usage and cost is on that single API.
OVERALL it does mean more usage for better models.
quote:
"Concluding Thoughts
Making Google Antigravity more accessible and affordable is just as important to us as the product advancements. While we would love to offer more generous quotas, the unprecedented demand for Antigravity and Gemini means we must balance our enthusiasm with real-world compute constraints. Ultimately, we hope these announcements give you the flexibility and clarity you need to focus more on what matters - experiencing liftoff."
source:
So many people abused gemini 3 flash in the past, i get it. But for a person who only have one antigravity account like me, that’s just unfair.
The X3 quota is meaningless if the AI itself nukes your quota by creating it’s own massive file to read continuously and eat your quota and lock you out for next 6 days!!
I was very careful to set up my workspace with only a few small text files and my .py files all combined about 150Kb. About 5 programs I am making with longest one only being 2000 lines.
I restricted permissions for the AI to not have access to any files or data outside of the workspace so it wouldn’t try reading any large files and use up the quota.
I have stayed well above using the 4 hour quotas ever since last night when locked out but apparently it doesn’t matter how careful you are apparently.
If there is not a large file for it to find and use up all your quota points on the AI will MAKE IT’S OWN file to eat your quota by reading endlessly!!!
Is this an example of poor coding on Googles part or is this perhaps by design to try forcing people to pay $200 a month? Inquiring minds want to know!
Both. Poor coding and force people to pay more.
As clearly stated in the notice, the quota increase and reset apply only to Gemini models:
"your Gemini quota has been reset... and increased by 3x"