Gemini CLI limits and the move to Antigravity CLI

What changed on June 18, 2026

On June 18, 2026, Login with Google stopped working in both Gemini CLI and the Gemini Code Assist IDE extensions if you're on Gemini Code Assist for individuals, Google AI Pro or Google AI Ultra. For those consumer and Google One tiers, Google replaced Gemini CLI with Antigravity CLI.

If your team uses Gemini Code Assist Standard or Enterprise, those plans follow a separate deprecation schedule. The gemini-cli repository stays open source under the Apache 2.0 license, and Gemini CLI keeps working with a paid Gemini API key, a Gemini Enterprise Agent Platform API key or Google Cloud.

Gemini CLI quotas by sign-in type

The table below shows the last Gemini CLI limits Google published per user per day (updated June 18, 2026), along with where each sign-in path stands today. Individuals who want to keep using Gemini CLI can switch to a paid Gemini API key.

• Requests are also limited per user per minute on top of the daily numbers.
• Run /stats model inside Gemini CLI to check your current usage.

Antigravity CLI quotas by plan

Google doesn't publish fixed request counts for Antigravity CLI, and your quota tracks the actual work the agent does rather than how many prompts you send. Antigravity shows your Weekly Limit Remaining and Five Hour Limit Remaining, and the Google AI Pro vs Ultra comparison covers what each plan includes.

Business teams buy Antigravity through Google Cloud: eligible Gemini Enterprise customers can pick Standard or Plus starting at $30 per seat per month, or a pay-as-you-go plan with a $0 seat fee and consumption-based pricing.

Going past the quota

• Google AI Pro and Ultra users can buy AI credits to keep working past their plan limits, billed at standard Gemini Enterprise consumption pricing.
• You control credit spending with the AI Credit Overages setting, which takes either Never or Always.
• Antigravity has no bring-your-own-key option and no custom endpoint setting.
• Running Gemini CLI with a paid Gemini API key bills per token on pay-as-you-go pricing instead of drawing from a subscription quota.

Antigravity CLI basics

• The CLI binary is agy.
• It's written in Go and runs on the same agent harness as Antigravity 2.0.
• You can extend it with Agent Skills, Hooks, Subagents and Extensions as plugins.
• Built-in slash commands include /agents to manage subagents, /boost for multi-agent deep reasoning, /codesearch to search your workspace and /context to inspect context usage.

For scripts or one-off commands, agy -p runs a single prompt in headless mode and prints the response to stdout, using the cached credentials from your first interactive agy sign-in.

# Sign in once with an interactive session
agy

# Then run a single prompt and print the answer
agy -p "In one sentence, what is a git rebase?"

Call your Google AI plan from your own code

If you want to call your Google AI Pro or Ultra subscription from your own scripts, servers or tools outside the terminal, you can connect your Google account on Plan2API and get a p2a_ API key for https://api.plan2api.com/v1. It speaks the standard OpenAI Chat Completions format (POST /v1/chat/completions) and exposes Gemini models through Antigravity, including gemini/gemini-4-argon-high and gemini/gemini-3.8-flash-high, each with a 1,048,576-token context window.

Every request counts against your Google plan's own usage limits, and Plan2API's Gemini subscription API adds no token quota of its own and never bills per token. If you connect more than one Google account, a request that hits a limit automatically fails over to the next account before any output is sent, and when every connected account is out of quota the API returns HTTP 429 with the code subscription_quota_exhausted and a Retry-After header in seconds.