Tool guide
Google Antigravity: The Antigravity IDE, Google Antigravity Pricing, and What the Free Tier Includes
Google's agentic coding platform, read from Google's own pricing page, plans doc and model matrix in August 2026. Including the quota figure Google does not publish and the model access that disappears on the enterprise plan.
No card to start · Review-first: the agent never merges
Pick a task
Plan
- planning
Files changed
Test run
Pull request
You review and merge. Agentcode never merges on its own.
The short answer
Google Antigravity is free to use and the free tier is the most generous in this category: $0 a month, every product feature, unlimited Tab completions, and a model list that includes Claude Opus 4.6 alongside the Gemini line. Above free there is no standalone Antigravity price at all. You buy a Google AI subscription, Pro at $19.99 a month or one of two Ultra plans at $100 and $200, and Antigravity quota comes with it. The catch is that Google publishes no absolute quota figure anywhere: the pricing page, the plans doc and the FAQ all describe limits in adjectives, and the only published numbers are the 5X and 20X Ultra multipliers, both measured against a Pro tier that has no stated number of its own. There is also one trap worth knowing before a company standardizes on it: the Claude and GPT-OSS models available on the free plan are not available on the Enterprise plan.
What is Google Antigravity?
Google Antigravity is Google's agentic coding platform. Google's own one-line description calls it a platform for orchestrating autonomous subagents, scheduled tasks and interactive code verification, which is a more accurate framing than "Google's answer to Cursor" even though that is how most people find it.
The first thing to get straight is that Antigravity is four products under one name, and a buyer who compares the wrong one against Cursor will reach the wrong conclusion. The editor is only one surface.
It is also the reason Gemini CLI is no longer sold to individuals. Google folded that product into this one and stopped serving Gemini CLI requests for free, Google AI Pro and Google AI Ultra users on June 18, 2026, with Antigravity CLI as the replacement. If you arrived here after that change broke something, Gemini CLI vs Claude Code covers what still runs, what a seat costs now, and how the comparison looks with the free tier removed.
Antigravity IDE
The full editor. Tab completion, natural-language Command edits, an agent side panel, and a browser the agent can drive against your running app so it can screenshot and verify what it built. This is the piece that competes with Cursor.
Antigravity 2.0
The Agent Manager: a command center for watching several background agents at once rather than one chat at a time. Google describes the editor as one surface among several, with the manager as the primary view.
Antigravity CLI
The terminal client, and the successor path for Gemini CLI users. Google documents it as preserving backward compatibility with the core constructs Gemini CLI popularized, and it still reads your existing GEMINI.md.
Antigravity SDK
A Python SDK for programming your own agents, tools, MCP integrations and scheduled tasks. There is no equivalent in most editor-shaped rivals, and it is the piece worth looking at if you want automation rather than assistance.
The browser control in the IDE is the genuinely distinctive part and it deserves more attention than the pricing does. The agent can drive a Chrome profile against your running application, take screenshots, and check that what it built actually renders. That closes a loop most coding tools leave open, where the agent writes something plausible and you find out later that the page is broken. If you want the fuller version of this, we wrote up what Google Antigravity is and who it is for separately.
Is Google Antigravity free?
Yes, and this is the part worth checking against anything else you have read, because the free tier has been rewritten more than once since launch and most write-ups are quoting a version of it that no longer exists. Here is what Google's pricing page and plans doc actually say today.
Swipe to see more
| On the free plan | What you get | Worth knowing |
|---|---|---|
| Price | $0 a month | Listed by Google as the Individual plan. No card, no trial clock. |
| Gemini models | Gemini 3.7 Flash, 3.6 Flash, 3.5 Flash, 3.1 Pro | The models doc shows the same Gemini line available on free as on Enterprise. |
| Third-party models | Claude Sonnet 4.6, Claude Opus 4.6, GPT-OSS-120b | Yes, really, on the free plan. And no, not on Enterprise. See the model matrix below. |
| Tab completions | Unlimited | Google's word, on the pricing page and again in the plans doc. |
| Command requests | Unlimited | Inline natural-language edits are not the thing that is metered. |
| Product features | All of them, including Scheduled Tasks and the CLI | Nothing is feature-gated behind a paid tier. The gate is quota, not capability. |
| Agent quota | "Meaningful quota", refreshed weekly | That phrase is the entire published specification. There is no number. |
Read the bottom two rows together and you have the whole shape of it: nothing is feature-gated, so the free plan is not a crippled demo. What is gated is agent quota, and the only published specification for that quota is the word "meaningful". If you are testing tools this week, install this one first, because it costs nothing and it is not a trial.
Google Antigravity pricing, tier by tier
Here is the structural fact that catches most people out: there is no Antigravity subscription to buy. Antigravity is bundled into Google's consumer AI plans, so the prices below are Google AI subscription prices that happen to raise your Antigravity quota. Google restructured those plans at I/O in May 2026, splitting Ultra into two tiers and cutting the top one from $250 to $200, which is why older articles quote figures that no longer match.
Swipe to see more
| Plan | Per month | Agent quota, as Google states it | What it means |
|---|---|---|---|
| Individual (free) | $0 | "Meaningful quota", refreshed weekly | Genuinely usable, and the most generous free tier in this category. The weekly refresh is the catch: spend it on Tuesday and you wait until the reset. |
| Google AI Pro | $19.99 | "High, generous quota", refreshed every five hours until the weekly limit | The five-hour refresh matters more than the size of the bucket. A bad morning stops costing you the week. |
| Google AI Ultra | $100 | 5X the Pro usage limit | New at I/O 2026. Google names Antigravity explicitly in the 5X claim, so this multiplier is about this product and not just the Gemini app. |
| Google AI Ultra (top) | $200 | 20X the Pro usage limit | Was $250 until I/O 2026. Same name as the $100 plan, which is a genuine source of confusion when someone quotes you an Ultra price. |
| Organization | Consumption | Billed through Google Cloud at Agent Platform pricing | Requires a Gemini Enterprise Standard, Plus or Pay-as-you-go edition. No seat price exists to quote: you are buying metered platform usage. |
Two footnotes on the numbers. The $100 and $200 Ultra figures and the 5X and 20X multipliers come from Google's own I/O 2026 announcement, which names Antigravity explicitly rather than talking only about the Gemini app, so those multipliers really are about this product. The $19.99 Pro price is the current US price but Google did not restate it in that announcement, so treat it as the market price rather than a quoted one. For the wider field, our AI coding tool pricing comparison puts every major tool's billing model side by side, and Cursor pricing breaks down the closest rival in full.
How Antigravity compares to Cursor, in short
Cursor is the right comparison because it is the same product shape: an editor with an agent in it, sold to individual developers first and teams second. The entry prices are within a cent of each other, so the decision is made everywhere except the sticker.
Swipe to see more
| Google Antigravity | Cursor | Edge | |
|---|---|---|---|
| Free tier | Full features, all models including Claude Opus 4.6, unlimited Tab, weekly agent quota | Hobby: limited agent requests and completions | Antigravity, clearly |
| Entry paid tier | $19.99 Google AI Pro | $20 Pro | Level |
| What the entry tier buys | A consumer Google bundle: storage, the Gemini app, and Antigravity quota | A coding tool, and $20 of third-party model usage | Depends what you want to pay for |
| Mid tier | $100 Ultra (5X Pro) | $60 Pro Plus ($70 of model usage) | Cursor, on price |
| Top tier | $200 Ultra (20X Pro) | $200 Ultra ($400 of model usage) | Level on price, not on clarity |
| Can you forecast the bill? | No. No absolute quota is published at any tier. | Yes. Included usage is stated in dollars. | Cursor, decisively |
| Team purchasing | No seat price. Gemini Enterprise edition, consumption billed. | $40 per user Standard, $120 Premium | Cursor for a small team, Google for a Cloud shop |
| Quota refresh | Weekly on free, every five hours on Pro and Ultra | Monthly allowance, then on-demand usage | Different in kind |
| Terminal client | Antigravity CLI, Gemini CLI compatible | Cursor CLI | Level |
| Agent SDK | Python SDK for custom agents and tools | No equivalent | Antigravity |
This is the summary version. The full tier-by-tier head-to-head, including what each vendor publishes and what people actually spend, is on Antigravity vs Cursor. The honest summary of that table: Antigravity wins on what you get for nothing, Cursor wins on what you can predict. If you are one developer choosing a tool for yourself this afternoon, the free tier makes the argument on its own. If you are the person who has to sign a purchase order and defend the number twelve months from now, a published dollar allowance is worth more than a generous unpublished one. Other pairs in this bracket: Windsurf vs Cursor, Cursor vs Claude Code, Cursor vs GitHub Copilot and Codex vs Cursor.
What decides whether this works for you
None of these five are on the pricing page, and between them they explain most of the disagreement you will find about this product online. Two of them are reasons to adopt it and three are reasons to check something before you do.
The product line moves fast, and the docs move with it
Re-checked against Google's own blog index on 2026-08-20. Antigravity 2.0, the Antigravity CLI, the Antigravity SDK and the Gemini Enterprise route were all announced together at I/O on 19 May 2026. Since then Google has shipped Custom Agents on 12 August 2026 and Gemini 3.7 Flash on 13 August 2026, and the pricing page now advertises the Google Cloud organization plan as available rather than coming. If you evaluated this product before May, you evaluated a different one.
Google publishes no quota number, anywhere
This is the single most important thing to understand before you standardize a team on it. The pricing page, the plans doc and the FAQ all describe quota in adjectives: "meaningful", "high, generous", "the highest, most generous". The only figures Google publishes are the 5X and 20X Ultra multipliers, and both are measured against Pro, which has no published number of its own. The ladder is 1x, 5x and 20x of a quantity nobody outside Google knows. You can work with that as an individual. It is a hard thing to put in a budget.
The model access inverts at the enterprise tier
Read the model matrix column by column and there is a genuine trap in it. Claude Sonnet 4.6, Claude Opus 4.6 and GPT-OSS-120b are available to free users, AI Plus, Pro and Ultra. They are not available on Enterprise. So a developer who has been happily running Opus on the free tier, whose company then buys Antigravity properly through Google Cloud, loses the model they were using. If your evaluation was done on a personal account, you did not evaluate the plan you are about to buy.
You cannot buy Antigravity on its own
There is no Antigravity subscription. As an individual you buy Google AI Pro or one of the two Ultra plans, which are consumer bundles carrying cloud storage, the Gemini app and other Google perks, and Antigravity quota rides along with them. As a company you buy a Gemini Enterprise edition and get metered Agent Platform billing through your Cloud account. Either way you are buying something larger than a coding tool, which is fine if you wanted the rest of it and wasteful if you did not.
Google AI Plus does not raise your Antigravity limits
Worth checking before you upgrade for the wrong reason. The plans doc lists exactly three quota tiers: Ultra, Pro, and one baseline tier covering "users not on AI Pro and Ultra plans". AI Plus sits in that baseline group, and the model matrix groups it with Free as a single column. If you buy AI Plus expecting more agent runs, you are buying storage.
The refresh cadence is the real difference between free and paid
Free quota refreshes weekly. Pro and Ultra refresh every five hours until you hit a weekly ceiling. That changes the texture of a working day more than the size of the allowance does. On the free tier, one afternoon of heavy agent use can leave you without an agent until the week turns over. On Pro, the same afternoon costs you a coffee break. If you are deciding whether the $19.99 is worth it, this is the reason, not the raw volume.
Who Google Antigravity is actually for
"Which is better" has no single answer here, because the two tools are strong in different places. Asked eight more specific ways, it does have answers.
Swipe to see more
| If this is you | Pick | Why |
|---|---|---|
| You want the best free AI coding tool available right now | Antigravity | Nothing else gives you Claude Opus 4.6, unlimited Tab and every product feature at $0. It is not close. |
| You already pay for Google AI Pro or Ultra | Antigravity | The quota is already included. Evaluating anything else without trying this first is leaving money on the table. |
| You need to forecast a monthly bill you can defend | Cursor | Included usage is published in dollars at every tier. Antigravity publishes no number at any tier. |
| You are buying for a team of five to fifty developers | Cursor | A per-seat price exists. Antigravity routes you through a Gemini Enterprise edition and consumption billing. |
| Your company already runs on Google Cloud with IAM and VPC Service Controls | Antigravity | The enterprise story is genuinely strong: your existing project, your billing account, your network controls, request and response logging. |
| Your team depends on Claude models specifically | Cursor | Antigravity Enterprise drops third-party model access. Check this before you migrate, not after. |
| You are migrating off Gemini CLI | Antigravity | The CLI is the documented successor path and keeps reading your existing GEMINI.md. |
| You want work to arrive as reviewed pull requests, not as an editor session | Neither, quite | Both are tools you sit inside and steer. See the section below on what they leave to you. |
Weighing a wider field than these two? Our best AI for coding roundup covers the rest of the category, and how to choose an AI coding assistant is the four-question version for people who have to justify the decision to somebody else.
What Antigravity and Cursor both leave to you
Antigravity is a good product and the free tier is a genuinely aggressive piece of pricing. But it is an editor, and an editor asks you to be there. You open it, you describe something, you watch the agent work, you steer it, you accept the diff. The agent manager makes running several at once easier, and Scheduled Tasks push further toward automation than most rivals go, but the center of gravity is still a person sitting in front of a window.
Agentcode is built for the other half of the job rather than to replace either one. You describe a task, it plans, edits the files, runs your test suite, and opens a reviewable pull request on the GitHub or GitLab repository you already have. It never merges on its own. The bill is $29 a month, flat, with no included-usage allowance to run down and no quota to wait out, which is why there is no row for us in the tables above: there is nothing to meter and nothing to forecast.
Where Antigravity genuinely beats us: the free tier costs nothing and ours does not, its Tab completion has no equivalent here, its browser-verification loop is better than anything we do in the editor, and if you already pay for Google AI Pro you have a capable agent sitting unused. If you want a frontier model at your fingertips while you type, install it today. If you want finished work arriving as pull requests on GitLab as well as GitHub, at a price you can forecast a year out, that is the case for us. Compare us directly against the closest rival on Cursor alternatives, or read what an AI coding agent does differently from an assistant.
Google Antigravity: your questions answered
Is Google Antigravity free?
Yes. Google lists an Individual plan at $0 a month with no trial clock. It includes every product feature, unlimited Tab completions, unlimited Command requests, the full Gemini model line, and third-party models including Claude Opus 4.6. The only limit is agent quota, which Google describes as a "meaningful quota" refreshed weekly.
How much does Google Antigravity cost?
There is no standalone Antigravity price. Free is $0. Above that you buy a Google AI subscription: AI Pro at $19.99 a month, or one of two AI Ultra plans at $100 and $200. Companies buy a Gemini Enterprise edition and are billed for consumption through Google Cloud.
What are the Google Antigravity rate limits?
Google does not publish an absolute number at any tier. The free tier gets a weekly-refreshing quota, Pro and Ultra refresh every five hours until a weekly ceiling, and the two Ultra plans are stated as 5X and 20X the Pro limit. Since Pro itself has no published figure, the multipliers are relative to an unknown.
What is the difference between Antigravity and Cursor?
Both are editors with an agent inside, and they cost roughly the same at the entry tier. The practical differences: Antigravity has a far more generous free tier, adds an agent manager and a Python SDK, and rides on a Google subscription. Cursor publishes its included usage in dollars, so you can forecast the bill.
Is Google Antigravity better than Cursor?
For a solo developer who wants the most capability for the least money, yes, mainly because the free tier is unusually good. For a team that needs a per-seat price and a predictable invoice, Cursor is the safer purchase today. Antigravity wins again if your company already runs on Google Cloud.
What models does Google Antigravity use?
Gemini 3.7 Flash, Gemini 3.6 Flash, Gemini 3.5 Flash and Gemini 3.1 Pro across every tier. Free, AI Plus, Pro and Ultra can also select Claude Sonnet 4.6, Claude Opus 4.6 and GPT-OSS-120b. Enterprise accounts cannot use those three. Image generation runs on Nano Banana 2 and is not user-selectable.
Is Antigravity replacing Gemini CLI?
Antigravity CLI is the documented successor path. Google states it preserves backward compatibility with the core developer-experience constructs Gemini CLI popularized, and it still reads your global GEMINI.md. Local skills move from .gemini/skills to .agents/skills, extensions are now called plugins, and MCP servers move into a standalone mcp_config.json.
Is Google Antigravity safe for enterprise code?
The enterprise controls are strong on paper. It requires a Gemini Enterprise edition, runs inside your own Google Cloud project, bills through your Cloud billing account, supports IAM role-based access, request and response logging, and VPC Service Controls. Google states enterprise prompts, responses, code and telemetry are never stored outside your private environments.
What is the Antigravity IDE?
It is the editor surface of Antigravity: a full development environment with Tab completion, natural-language Command edits, an agent side panel, and a browser the agent can drive against your running application to screenshot and verify its own work. The CLI, the agent manager and the Python SDK are separate surfaces of the same product.
Does Google Antigravity work with existing VS Code extensions?
The IDE is a full editor with plugin support, and Google documents plugins, hooks, rules, workflows and MCP servers as first-class features. Google does not state which extension marketplace it draws from, so if a specific extension is load-bearing for your team, verify it directly before you migrate rather than assuming parity.
Last updated: August 2026
Still comparing?
Full pricing breakdowns for each tool: Cursor pricing, Claude Code pricing, GitHub Copilot pricing, Devin pricing and Kiro pricing.
Or read what Google Antigravity is, Codex vs Claude Code, what agentic coding means, or step back to AI coding tool pricing across every tool.
A flat bill, on your repo
Connect a GitHub or GitLab repository you already have, describe a task, and read the pull request it opens. No included allowance to run down and no quota to watch.