Hello everyone,
I’ve been using Google Antigravity heavily for managing my agent workflows, but recently I’ve noticed a major shift in how long my token allocations and AI credits actually last.
It feels like my baseline quota is burning out almost instantly compared to a couple of weeks ago. I know under the hood that rate limits are correlated with how much work an agent does per task, but even on relatively straightforward multi-step workflows, my credit balance drops massively.
With the shift toward running multiple subagents in parallel, it feels like the token consumption has multiplied aggressively. Is anyone else noticing their flexible credit pools disappearing faster than usual, or have there been recent behind-the-scenes adjustments to how consumption pricing is calculated for complex agent loops?
I’d love to know if people are finding ways to optimize their agent behavior files to keep costs down, or if upgrading to the higher tiers is the only viable workaround now.
Thanks!