Hey everyone,
I’ve been using Antigravity for about 4 months now and have an active Pro subscription. I intentionally stayed on the IDE build rather than moving to the standalone app. Here are my specs:
-
Antigravity Version: 1.23.2
-
VSCode OSS Version: 1.107.0 (user setup)
-
Commit: 15487b3041e65228cae24980a3f796c905ef582c
-
Date: 2026-04-16T08:28:19.366Z
-
OS: Windows_NT x64 10.0.19045
The Issue: Recently, right around the time Flash upgraded to 3.5, my AI quota doesn’t seem to be getting utilized at all. Even when I explicitly select Gemini 3.1 Pro from the model dropdown, it feels like my entire workspace is being silently powered by Flash.
I am currently building out the core mathematical logic. I am focusing entirely on generating complex scripts. The code I need requires deep architectural reasoning, but the AI’s outputs are far less sharp than I expect from Pro seems more like Flash . The responses are generating instantly, but they fail to handle the heavy mathematical lifting for the scripts.
Through all of this, my AI quota just stays exactly as it is and doesn’t drain.
Has anyone else experienced this silent fallback on 1.23.2? Is there a known way to force the system to actually use Pro for the heavy logic without it offloading to Flash?
Any help would be appreciated!