Hi Google Antigravity Product Team,
I am a $200/month tier subscriber using Google Antigravity for real-world agentic tasks. I am writing to strongly request the removal of the strict 5-hour rolling quota cap on Claude Opus and similar frontier models.
Current Pain Point
Complex agentic tasks require sustained multi-step reasoning and tool execution. Right now, the 5-hour quota cap frequently triggers mid-task before the agent can finish its work. When this happens:
- Tasks fail mid-execution with zero usable output: The in-progress run stalls or aborts right before completion, leaving the task unfinished and unusable.
- Consumed quota is completely wasted: All the quota already spent on that unfinished run goes down the drain with nothing to show for it.
- Unpredictable reliability on a $200/month plan: Even as a high-tier paying subscriber, hitting hard 5-hour walls makes Opus unreliable for serious, multi-step engineering and research workflows.
Strong Request
- Remove the 5-hour hard cap on Opus and other frontier models for $200/month subscribers (allocate usage from a monthly or weekly pool rather than enforcing short 5-hour rolling windows).
- At minimum, provide mid-task completion protection or burst overflow: Never cut off an active agent run mid-task so that already-consumed quota is wasted without delivering a result.
As a $200/month subscriber, predictable end-to-end task completion is essential. Please reconsider and remove the 5-hour quota cap on Opus.
Thank you!