Reviewed by Jonathan West · Updated Nov 1, 2026

Google Antigravity Pricing in 2026: Free Plan to $200 Ultra

What Google's agent-first coding platform costs today, and the quota limits teams should plan around

Reviewed by Jonathan West · Updated Nov 1, 2026

Yes, Google Antigravity has a free Individual plan at $0 per month with basic weekly limits. Under the plan updates announced on May 19, 2026, individual paid tiers cost $20 per month for Google AI Pro and $100 or $200 per month for Google AI Ultra. For organizations purchasing through Google Cloud, Business seats start from $30 per seat per month.

Launched in November 2025, Antigravity is an agent-first coding platform built on VS Code. It supports multiple AI models, including Google's Gemini and third-party options, and offers browser automation and multi-agent orchestration. See Cursor vs. Google Antigravity for details on the models it supports.

Free access does not mean unlimited usage, since Google refreshes free allowances weekly while paid quotas reset every five hours up to a weekly cap. Google publishes no numeric limits for any plan, and extra usage is billed as overage AI credits only when the AI Credit Overages setting is Always.


Google Antigravity Plan Tiers and Quotas

Google Antigravity offers an Individuals plan at $0 per month, Google AI Pro at $20 per month, two Google AI Ultra tiers at $100 and $200 per month, and Business access through Google Cloud starting from $30 per seat per month or via pay-as-you-go billing. The free Individuals tier provides unlimited Tab completions and Command requests under basic weekly limits. Subscribers on the $20 Pro plan receive a higher quota that refreshes every five hours until reaching a weekly cap, a cadence worth comparing in Cursor vs Google Antigravity.

According to the Google Antigravity plan update, the $100 Ultra tier provides five times the token rate limit of Pro, while the $200 Ultra plan supplies twenty times that baseline. Both Ultra tiers refresh quotas every five hours, drawing from a unified Gemini pool based on Application Programming Interface (API) rates. While non-Gemini models keep a separate fixed rate limit, developers can compare these allowances against other environments in Google Antigravity vs Claude Code.

Base plans do not include credits, leaving purchased credits solely for overages managed by setting credit draw to Never or Always. For eligible Gemini Enterprise customers, the Google Antigravity pricing page outlines Business options starting at $30 per seat monthly alongside a $0 seat fee pay-as-you-go tier with spend caps. As stated in the Google Antigravity plans documentation, quotas depend on system capacity and task complexity, meaning Google may modify these tiers over time.


Is Google Antigravity Free?

Google Antigravity provides a free Individual plan at $0 per month with basic weekly rate limits, unlimited Tab completions, and unlimited Command requests. Its quota refreshes on a weekly schedule. Google does not publish specific numeric quotas for free accounts.

Free accounts include access to Gemini 3.8 Flash, Claude Sonnet 4.6, Claude Opus 4.6, and GPT-OSS-120B according to the Google Antigravity model documentation. Google has scheduled Claude Sonnet 4.6, Claude Opus 4.6, and GPT-OSS-120B for removal from Antigravity on November 2, 2026.

Gemini 4 Argon is restricted to Google AI Ultra and eligible Gemini Enterprise accounts, while Claude Sonnet 5.5 and Claude Opus 5.5 require a non-trial Google AI Pro plan or higher. Free accounts also omit enterprise options like Gemini 3.1 Pro, Gemini 3.6 Flash, and Gemini 3.7 Flash. Our guide explains who qualifies for a Google AI Pro free trial.

The free tier works across the Antigravity 2.0 desktop application, the Antigravity integrated development environment (IDE), supported third-party editor extensions, and the agy command-line interface (CLI) covered in our Antigravity CLI guide. Developers who exhaust the basic weekly allowance or need a faster 5-hour refresh cycle should upgrade to Google AI Pro for $20 per month. You can review the latest plan specifications directly on the Google Antigravity pricing page.

  • $0 per month Individual tier with basic weekly rate limits.
  • Unlimited Tab code completions and Command requests.
  • Access to Gemini 3.8 Flash, with Claude 4.6 variants departing November 2, 2026.
  • Functional across the 2.0 app, the IDE, extensions, and the agy CLI tool.
  • Quotas refresh once per week instead of every 5 hours.
The Individual plan is a $0 tier with unpublished weekly usage quotas, and Google publishes no separate student discount or student tier.

How Antigravity Usage Quotas Work

Google Antigravity usage quotas reset on distinct schedules depending on the account tier. The Individual free tier refreshes its quota weekly, while Google AI Pro and Google AI Ultra refresh every 5 hours until reaching an overall weekly limit. Ultra at $100 per month provides 5 times the token rate limit of Pro, and the top $200 per month Ultra tier raises that allocation to 20 times Pro.

Model selection determines how quickly an active session drains that capacity. Requests to Gemini Flash and Gemini Pro share one combined quota drawn down according to API pricing, whereas non-Gemini models operate against a separate fixed limit.

Google publishes no numeric quotas for these tiers on the official pricing page or in the plan documentation. When an account exhausts its allocation, the AI Credit Overages setting determines whether the agent halts or keeps running. Setting overages to Always bills continued work at Gemini Enterprise consumption pricing, while Never stops requests until the quota refreshes. Google Antigravity does not support bringing your own API key for extra limits.

Two documented problems have hit accounts at the quota limit. Google acknowledged a January 2026 Pro plan bug reported on the developer forum where five-hour refresh cycles failed and trapped accounts in multi-day lockouts. Agents can also crash with HTTP 429 RESOURCE_EXHAUSTED errors mid-run when rate boundaries hit unexpectedly.

Engineering leads should keep a fallback coding agent ready, such as Claude Code or Cursor, if a session runs into a hard rate wall.

Set AI Credit Overages to Always for production pipelines, or maintain a secondary tool to avoid stalling when Google Antigravity hits its weekly token ceiling.


Security and Data Handling for Teams

For business use, the key question is how Antigravity handles your code and prompts, because it sends context to AI models to do its work.

It can route work to multiple models, including third-party ones. That means your data-handling review must cover each model provider you enable.

Antigravity is included in eligible Gemini Enterprise Standard and Plus tiers as of August 20, 2026, according to Google's enterprise announcement. That path adds admin budget caps, capped overages, audit logging, sandbox rules, and Model Context Protocol (MCP) restrictions. Individual Google AI Pro and Google AI Ultra accounts do not provide those centralized governance controls.

Regulated teams should route rollouts through Gemini Enterprise. Keep turbo-style auto-run turned off and require explicit approval on shell actions, because an automated Turbo-mode command wiped a user's drive in December 2025, as reported by The Register. Pillar Security discovered a prompt-injection flaw leading to code execution even in strict mode, which Google patched on February 28, 2026, and disclosed in April 2026, per CyberScoop. Keep agent access restricted to the project folder and do not feed it untrusted web content.

On training use specifically: Google's Antigravity terms page (antigravity.google/terms) does not spell out a clear, Antigravity-specific commitment on whether code and prompts are used to train models. Do not assume a no-training guarantee. Before sending code you would not want used for training, get that commitment in writing in a Gemini Enterprise contract.

Regulated teams should mandate Gemini Enterprise seats, require manual confirmation on every terminal command, and restrict agent access to designated project folders.

Is Google Antigravity Worth Paying For?

The free plan is worth adopting for individual developers evaluating agent-first coding. Google AI Pro at $20 per month is the practical step up for daily development and access to Claude 5.5 models. Google AI Ultra at $100 or $200 per month suits heavy developers who regularly hit five-hour quota limits on Pro.

Google provides Gemini 4 Argon to Ultra and eligible Gemini Enterprise accounts in the Antigravity 2.0 desktop app and Antigravity CLI, but omits the model from the Antigravity IDE. Teams that need administrative oversight should purchase Gemini Enterprise rather than individual accounts. Gemini Enterprise provides admin budget caps, audit logging, and sandbox restrictions.

Antigravity is not the right choice for engineering teams that need published numeric quotas for capacity planning. If Google publishes token counts or request limits on its pricing page or plans documentation, we would revisit that verdict. Layer3 Labs can help your team structure and evaluate an internal pilot across developer workflows.

Start on the free plan, track when you hit the weekly cap, then decide between Pro and Ultra based on your reset patterns.


What Antigravity Is For, and How It Compares With Cursor, Claude Code and Codex

With Google Antigravity, developers supervise agents that work asynchronously across the code editor, terminal, and browser. Google Antigravity's Agent Manager lets you spawn parallel runs that produce plans, task lists, and browser screenshots for verification. At its November 2025 launch, Google Antigravity led with Gemini 3 Pro, with support for Claude Sonnet 4.5 and OpenAI GPT-OSS.

Choosing among Google Antigravity, Cursor, Claude Code, and OpenAI Codex comes down to cost certainty and team management controls. Cursor and Claude Code offer published seat pricing with single sign-on (SSO) and administrative oversight. On its pricing page, Google Antigravity lists a $0 Individual plan, Pro at $20, Ultra at $100 or $200, and Business seats from $30. Plan spend is predictable, but Google publishes no numeric quotas, so teams cannot size capacity from the published plans.

Review our evaluation of the best AI coding tools for business to see how these architectures fit broader engineering rollouts.

  • Cursor: Fits teams wanting a mature editor where you approve edits. Pricing sits at $20 monthly for Pro and $40 per user per month for Teams, backed by team Privacy Mode and audit logs on Enterprise. Read our detailed Cursor vs Google Antigravity guide.
  • Claude Code: Fits terminal-first developers who want careful, review-heavy execution. Anthropic includes it on paid Claude plans starting at $20 per month, with team administration on higher tiers and no model training on data in paid plans. Check our Claude Code pricing guide and Google Antigravity vs Claude Code review.
  • OpenAI Codex: Fits automated repository work where agents inspect files, run test suites, and open pull requests using GPT-5.6. Compare both agent environments in our OpenAI Codex vs Google Antigravity analysis.
  • Google Antigravity: Fits teams testing multi-agent workflows across paid tiers, with administrative budget caps and audit logging available through Gemini Enterprise since August 20, 2026. The remaining planning gap for engineering budgets is that Google publishes no numeric quotas for any tier.

Frequently Asked Questions

  • Yes, Google Antigravity offers a free Individual plan that includes basic weekly rate limits alongside unlimited Tab completions and Command requests. Developers who need higher limits can upgrade to Google AI Pro for $20 per month, Google AI Ultra for $100 or $200 per month, or Google Cloud Business seats starting at $30 per seat per month. You can review the full breakdown of features across each plan on the Google Antigravity pricing page.
  • Google Antigravity limits usage through reset intervals rather than published token numbers, refreshing free accounts weekly while Pro and Ultra reset every five hours up to a weekly cap. Google AI Ultra expands that allowance to five times Pro's rate limit at $100 per month, or twenty times at $200 per month. Gemini models share a single quota drawn down at application programming interface (API) pricing, whereas non-Gemini models follow a separate fixed limit. Because Google publishes no exact numeric quotas, verify plan details on the Antigravity plans page.
  • Your "AI Credit Overages" setting in Google Antigravity decides what happens. Set to Never, requests pause until the quota refreshes. That takes 5 hours on Google AI Pro and Ultra, or one week on the free tier. A paid account that reaches its weekly cap waits for the weekly refresh. Set to Always, you keep working on overage credits billed at Gemini Enterprise consumption pricing. Google acknowledged a quota reset bug in January 2026 that trapped some Pro accounts in multi-day lockout loops. If a 5-hour reset does not arrive, switch overages to Always or report it on that forum thread.
  • Google Antigravity is free on the Individual plan at $0 per month. Paid plans are Google AI Pro at $20 per month and Google AI Ultra at $100 or $200 per month. Organizations can buy Google Cloud Business seats from $30 per seat per month or use pay-as-you-go consumption billing. Eligible Gemini Enterprise Standard and Plus editions include Google Antigravity, and the Google Antigravity pricing page lists each tier.
  • Google does not publish credit-pack prices. AI credits are overage-only since May 19, 2026, billed at Gemini Enterprise consumption pricing. You manage this extra usage through the AI Credit Overages setting by selecting Never or Always. Antigravity does not support bringing your own API key for added limits, so check the Google Antigravity plans documentation to review your team's consumption billing options.
  • Google Antigravity can be used for business code if your team deploys it through Gemini Enterprise, which adds audit logging, admin budget caps, and sandbox and MCP restrictions. Because Antigravity also includes third-party models like Claude and GPT-OSS alongside Gemini, your data review must cover each model provider you enable. Pillar Security uncovered a prompt-injection flaw that allowed code execution even in strict mode before Google fixed it on February 28, 2026, as reported by CyberScoop.
  • Yes. Since August 20, 2026, Google Antigravity is included in eligible Gemini Enterprise Standard and Plus, per Google's enterprise announcement. Those tiers include admin budget caps, capped overages, audit logging, and sandbox and MCP restrictions. Individual Google AI Pro and Ultra plans omit management controls. Business seats are also available through Google Cloud from $30 per seat per month, or pay-as-you-go with consumption pricing.
  • Google's Antigravity terms page does not spell out a clear, Antigravity-specific commitment on whether code and prompts are used for model training. Do not assume a no-training guarantee. Get a written no-training commitment in a Gemini Enterprise contract before sending code you would not want used for training.
  • Pilot Google Antigravity on the free plan, then move to Google AI Pro, Ultra, or Gemini Enterprise seats once the agents earn the spend. Google publishes no numeric usage quotas on its pricing page. We advise keeping a second assistant like Cursor or Claude Code ready for deadline work.

Thinking about piloting Google Antigravity?

We help you pilot Antigravity on the free plan, set data-handling guardrails, and decide between Pro, Ultra and Gemini Enterprise seats before you standardize.

Book Your Free AI Workflow Audit