Gemini 3 Pro cant finish even one task without reaching quota limit. 3 days earlier I was able to prompt certain big tasks without issues. In summary there were 6 big tasks which it handled extremely well before reaching quota limit. Now it cant event complete one big task and I have to wait after 1 prompt to prompt it again. Thats very annoying.
I even bought AI Pro plan and still nothing changed
I use Antigravity
Google AI Studio works as usual.
I see many posts about all the problems people are having with Gemini right now because its not usable anymore due to immense token usage.
This happens in a current project which I am still on after a couple of days. It uses chat history data and muliple checkpoint,todo files.
Please fix this asap because this issue makes Gemini useless.
Hello @obyondx, thank you for sharing your experience with Gemini 3 Pro.
To help optimize your token usage, I recommend looking into the following three configurations:
Adjust the Thinking Level: Gemini 3 Pro and Flash utilize an internal chain-of-thought process that consumes tokens before generating an answer. For routine tasks, setting this to low can significantly reduce both cost and latency. Learn more about Thinking here.
Utilize Context Caching: If your workflow involves sending the same large files (such as PDFs or long codebases) repeatedly, avoid sending them in every request. Context Caching is much more efficient for this use case. View Caching documentation.
Lower Media Resolution: High-resolution analysis for images and video is token-intensive. Unless you specifically need to read fine print (OCR), lowering the resolution is a great way to save tokens.