I would like clarification regarding Antigravity’s quota behavior because the current UI appears inconsistent from a user perspective.
Observed Behavior:
1. I used Gemini 3.5 Flash (Medium) for coding tasks.
2. After some usage, quota warnings appeared.
3. I switched to Claude Sonnet 4.6.
4. After additional usage, Claude Sonnet also became restricted.
5. When opening the model selector, warning indicators appeared across:
* Gemini Flash variants
* Gemini Pro variants
* Claude Sonnet
* Claude Opus
* GPT-OSS 120B
The confusing part is that GPT-OSS had not been used directly during the session.
At the same time, the interface reported:
“Individual quota reached.”
Question:
If quotas are individual per model, why are multiple unrelated model families showing warning indicators simultaneously?
If quotas are shared across models, could the UI communicate this more clearly?
Current Confusion:
Users cannot easily determine:
* Whether quotas are model-specific
* Whether quotas are shared
* Whether limits are account-wide
* Whether warning indicators represent shared pool exhaustion
* Why unused models become restricted
Suggestion:
A quota dashboard showing:
* Shared quota remaining
* Per-model usage
* Weekly usage
* Session usage
* Clear explanation of warning indicators
would significantly improve transparency.
I have attached a screenshot showing simultaneous warning indicators across Gemini, Claude, and GPT-OSS models.
This is not primarily a request for higher quotas.
It is a request for clearer quota visibility and explanation of model availability behavior.