Codex vs Claude Code vs Antigravity: plans, usage limits and prices compared (2026)

Short answer. On published prices, Claude Code and Codex both start paid use at $20 a month ($20 for Claude Pro, $20 for ChatGPT Plus), and both have $100-and-up tiers for heavy use. Antigravity is the only one with a free plan that lists unlimited Tab completions, but the price of its paid Google AI Pro and Ultra tiers was not shown on the pages we checked. Only Codex publishes concrete message ranges per five hours; the other two describe usage in multiples or in words, so you cannot compare how much work each plan buys from the pages alone.

OpenAI Codex, Anthropic’s Claude Code and Google Antigravity are AI coding agents that read a project, edit files and run tasks. They are sold differently: Codex and Claude Code come inside chat subscriptions (ChatGPT and Claude), and Antigravity is bundled with Google’s AI plans. This comparison uses only the vendors’ official pages, checked on November 11, 2026. Plans, models and limits change often, so confirm before you buy.

At a glance

	Codex (OpenAI)	Claude Code (Anthropic)	Antigravity (Google)
Free plan	Yes: Free ($0) and Go ($8); limits not stated	Claude Free does not include Claude Code	Yes: $0, unlimited Tab completions and Command requests, “basic weekly” limits
Entry paid plan	Plus: $20/month	Pro: $20/month ($17/month billed annually)	Google AI Pro: price not shown
Heavier plans	Pro from $100/month ($100, $200 or $500)	Max $100/month (5x Pro); Ultra $200/month (10x Pro weekly)	Google AI Ultra: price not shown
Team plans	Business: $20/user/month annually, $25 monthly	Team: Standard $25/seat, Premium $125/seat (annual $20 and $100)	Business via Google Cloud: from $30/seat/month; pay-as-you-go option
How usage is expressed	Local messages per 5 hours for Plus (e.g. 15-160 for GPT-6.1 Sol) plus weekly limits	Multiples of Pro (Max 5x) for 5-hour and weekly limits	Quota “refreshed every five hours” on Pro and Ultra, weekly on others; no numbers
Extra usage	Buy credits (Plus and Pro)	Usage credits (Pro and above)	AI credit overages (Pro and Ultra)
API-key option	Yes, pay at API pricing	Enterprise: usage at API rates	Not offered (no bring-your-own-key)

What Codex offers

Per OpenAI’s pricing page, Codex is included in ChatGPT plans: Free ($0), Go ($8/month), Plus ($20/month), Pro (from $100/month, with tiers at $100, $200 and $500), Business ($20 per user per month billed annually, or $25 monthly) and Enterprise/Edu (contact sales). There is also an API-key option that pays for Codex usage at API pricing.

Models. Free and Go list GPT-6 Luna in the desktop app, subject to rollout. Plus and Pro add GPT-6.1 Sol and others, and the $500 Pro plan adds an Ultrafast mode.
Limits. For Plus and standard Business, the page gives local-message estimates per five-hour period that vary by model, for example 15 to 160 for GPT-6.1 Sol and 350 to 3,000 for GPT-6 Luna. Cloud tasks can use more of the allowance than local messages, and weekly limits may apply. Pro lists no five-hour limit currently, with weekly limits not stated. Free and Go limits are not stated.
Extra usage. Plus and Pro users can buy credits after reaching limits. The page lists credit rates per million tokens by model (for example GPT-6.1 Sol: 50 input, 2.5 cached input and 250 output).

What Claude Code offers

Per Anthropic’s pricing page, Claude Code is included on paid Claude plans and not on Free.

Plan	Price	Usage (relative to Pro)
Pro	$20/month, or $17/month billed annually	1x for 5-hour and weekly limits
Max	$100/month	5x
Ultra	$200/month	10x Pro weekly; no 5-hour limit for verified accounts
Team	Standard $25/seat (annual $20); Premium $125/seat (annual $100)	Premium is 5x Standard weekly
Enterprise	$20/seat/month, 20-seat minimum, plus usage at API rates	Spend limits set by admins

The page says usage limits depend on the model, chat length and file size. Max and Ultra include API credits ($100 and $200 per month). Which models run inside Claude Code is not stated; the page lists model access by plan (Pro includes Opus, Sonnet and Haiku, with the newest model available through usage credits).

What Antigravity offers

Per Google’s pages, the Individuals plan is free and lists access to “a variety of agent models,” unlimited Tab completions, unlimited Command requests and “basic weekly” rate limits. Gemini models are the core agent models (the docs name Gemini 3.8 Flash), and all plans list access to features such as scheduled tasks and the CLI.

Google AI Pro and Ultra. Antigravity is included in both, with a higher quota “refreshed every five hours until the weekly limit is reached” on Pro and “the highest, most generous quota” on Ultra. Ultra also lists access to third-party models; for Pro, third-party models are not stated. The pages we checked did not show the monthly prices of these Google subscriptions.
Credits. Pro and Ultra can use purchased or promotional AI credits for overage. An “AI Credit Overages” setting can be set to Never or Always. Credit prices and pool sizes are not stated.
Business. Standard and Plus plans via Google Cloud start at $30 per seat per month, with a pay-as-you-go option that has a $0 seat fee.
Not offered. Bring-your-own-key for extra rate limits.

Where they differ

Free tier. Antigravity’s free plan lists unlimited Tab completions and Command requests. Claude Free excludes Claude Code. Codex lists Free and Go access but no limits for them.
Price transparency. Codex and Claude Code publish every subscription price. Antigravity’s paid tiers depend on Google AI Pro and Ultra, whose prices did not appear on the pages we checked.
How limits are described. Codex gives message ranges. Claude Code gives multiples of Pro. Antigravity gives qualitative words (“generous,” “highest”). Only the first lets you estimate volume, and the ranges depend on the model you pick.
Model choice. Codex uses OpenAI models, Claude Code lists Anthropic models by plan, and Antigravity is built around Gemini, with third-party models listed only for Ultra.
API route. Codex offers a plain API-key option. Antigravity says it has no bring-your-own-key option. Claude Code’s page mentions API-rate usage for Enterprise.

Who should choose which

You want to try an agent for free first: Antigravity’s free plan lists unlimited Tab completions; check its weekly limits against your workload.
You want the entry paid price you can read: Codex Plus and Claude Pro are both $20 a month.
You want published usage estimates: Codex is the only one with message ranges, per model.
You work heavily and want a higher tier: Claude Max ($100, 5x Pro) and Codex Pro (from $100) have published prices; Google AI Ultra’s price was not shown.
You already pay for ChatGPT, Claude or Google AI: check whether the agent you need is already included in that plan before buying a second one.

If you are weighing editor-based tools instead, see Cursor vs GitHub Copilot pricing.

What we couldn’t verify

How much real coding work each plan’s limit allows. The vendors’ units (messages, multiples, quota) are not convertible.
The monthly prices of Google AI Pro and Ultra, and Antigravity credit prices.
Free and Go limits for Codex, Team five-hour limits for Claude, and Codex Pro weekly limits.
Which Claude models run inside Claude Code on each plan.
Regional or currency differences. All prices are as shown in US dollars.

FAQ

Is Claude Code free? No. Anthropic’s pricing page marks Claude Code as not included on the Free plan; it starts with Pro at $20 per month.

Is Codex included in ChatGPT Plus? Yes. OpenAI’s pricing page lists Codex in ChatGPT Plus at $20 per month, with local-message estimates per five hours that vary by model.

Does Antigravity have a free plan? Yes. Google’s pricing page lists an Individuals plan at $0 with unlimited Tab completions, unlimited Command requests and basic weekly rate limits.