I signed up for an API key (the paid tier, not the free tier; I think that matters after going through the documentation on paid models with a free tier) after I trialed Gemini CLI for a week using Google OAuth (i.e. all free usage, I think it’s called Gemini Code Assist).
I assumed the models available when using Antigravity CLI with an API key would be the models listed here: https://ai.google.dev/gemini-api/docs/pricing Was that an incorrect assumption? (it’s possible I’m looking at the wrong pricing page (there’s a few floating around).
When authenticating with Antigravity CLI using my API key and cloud project ID, I only have the 3.x models available. How do I switch to the 2.x models in Antigravity CLI? All my agent scaffolding is built on 2.x models for grunt work and reasoning/planning hand-off to 3.x models. If I can’t select a 2.x model in the CLI, how do I know I can call a 2.x model in my code snippets?
I have included labeled screenshots of model availability in Antigravity CLI when authenticating with OAuth vs API key/Cloud Project ID. I have also included screenshots of model availability when authenticating in Gemini CLI with OAuth versus API key/Cloud Project ID.
From the screenshots you can see I have full choice of the 2.5 models in Gemini CLI with my API key.
Question: Is the intention we keep using Gemini CLI until the full models are available in Antigravity CLI?
Noting that Gemini CLI will be unalived on 18 June: An important update: Transitioning Gemini CLI to Antigravity CLI - Google Developers Blog
I don’t use any of Antigravity IDE, Antigravity web, AI Studio. All my agent skills, hooks, and subagents are read in-environment by the CLI with AGENTS.md files.
