Google Antigravity Adds Gemini 3.5 Flash (Low), Cuts Token Usage 45%, Resets All Quotas

According to Beating, Google Antigravity announced recently that it has launched Gemini 3.5 Flash (Low) model to reduce token consumption on simple tasks. The new model cuts token generation by approximately 45% compared to Gemini 3.5 Flash (Medium), and outperforms the previous-generation Gemini 3 Flash (High) on software engineering tasks, according to internal testing data.

To ensure developers have sufficient resources over the coming week, Antigravity has reset Gemini quotas across all plans.