Google AI Studio Pricing (2026): Free Tier, Plans & API Costs

Wondering whether Google AI Studio is free, or when you'll need to pay? This guide breaks down the plans, explains the usage limits, and shows which costs can appear when you build and publish an app.

Google AI Studio gives you room to experiment for free. The confusing part is figuring out what happens when you need more access or want to publish your app. A Google AI subscription, Gemini API usage, and hosting each have their own pricing, so paying for one doesn't necessarily cover the others.

You don't need to sort all of that out before trying your first idea. The Free plan is a good starting point, and you can decide whether to pay once you know what you want to build and where you're running into limits.

Is Google AI Studio free?

Yes, Google AI Studio has a Free plan that lets you get started without buying a subscription. There are limits on usage and model access, and Google offers paid plans for people who need more.

Where it gets a little confusing is the difference between building an app and running it. You might create a form for free, for example, then add an AI feature that makes a paid API request whenever someone submits it. The table below separates those costs.

Google's AI Studio plan guide distinguishes subscription benefits from direct API billing. The Gemini API pricing page lists which models have a free tier.

Google AI Studio Free vs Pro vs Ultra

If you're learning how AI Studio works or trying out an app idea, we'd start with Free. Pro becomes worth considering when you keep running out of access while building. Ultra is a bigger commitment, so it makes sense to understand what the lower plan is missing before paying for it.

Pro gives you higher quotas, while Ultra offers the highest quotas for supported Studio use. Those benefits apply to the Playground and supported Build features. If you're using the Gemini API in a separate app, though, that usage still has its own billing.

The US monthly subscription prices checked on October 23, 2026 are:

These are the monthly subscription prices on Google's US subscription page, rather than the cost of hosting an individual app. Your local checkout will show any taxes, offers and renewal terms. You'll also see usage multipliers advertised for the Gemini app; those aren't a reliable way to calculate your AI Studio or API allowance.

The credit names can be confusing, too. Google One AI credits aren't supported in AI Studio, while Google Cloud credits are a separate system. Some eligible subscribers can receive Cloud credits through the Google Developer Program, subject to its requirements.

Before upgrading, it's worth checking which allowance you're actually using up. Running out of access while editing in Studio is different from a published app hitting its API quota. A higher Studio subscription may help with the first problem without solving the second.

What are the free-tier limits?

The answer depends on the model and your account. That's why you may see people reporting different numbers of free prompts. The most useful figure is the allowance shown for the model you're using in your own account.

For the API, limits can include requests per minute, input tokens per minute and requests per day. Hitting any one can stop a request even if you have room under the others. Daily API request quotas reset at midnight Pacific time, and requests from multiple keys in the same project share project limits.

For example, your app could work all morning, then start returning errors when several people try it at once. You may still have plenty of daily allowance left, but too many requests are arriving within the same minute.

Open your active rate limits in AI Studio, check the relevant model, and compare the error with the limit shown. Google's rate-limit documentation explains the different measurements.

Because those limits are shared, creating another API key in the same project won't give you a fresh allowance. Depending on the error, you may be able to solve the problem by spacing out requests or removing unnecessary calls before moving to a paid tier.

When does Google AI Studio start costing money?

There are a few ways you can start paying: subscribing for more access in Studio, enabling paid API usage, or using hosting and other services for your app. It's possible to pay for one without needing all three.

When you set up paid API billing, Google can ask for a minimum $5 prepayment, although some accounts have postpay options. That money goes toward usage. It doesn't buy a month of unlimited requests.

If you're unsure whether a key is using free or paid access, check the project it belongs to. You'll find its tier in the Billing Tier column on the Projects page, with payment settings and balances on the Billing page. Keys inherit the project's billing status, so a newly created key isn't necessarily a free one.

You can also set a project spend cap where the feature is available. Google currently labels it experimental, and delayed billing data or long-running tasks can produce overages, so leave some room in your budget. Google's billing guide explains how it works.

For a first paid project, try a small batch of realistic requests and look at the usage afterward. That gives you a much better starting point for a budget than guessing how many tokens each visitor will use.

Gemini API pricing: a worked example

If you haven't used token-based pricing before, the main distinction is input versus output. You pay for what you send to the model and for what it generates. Where applicable, the model's thinking tokens also count toward output charges.

As of October 23, 2026, Google's Standard paid rates include:

The Flash rates above apply through December 31, 2026, then become $1.50 input and $7.50 output. The Argon rates apply through November 30, 2026, then become $4 input and $20 output. These are USD API rates, not subscription prices. Verify the current rate card before budgeting.

What would 1,000 AI requests cost?

Suppose a small app makes 1,000 requests, each using 2,000 input tokens and 500 billable output tokens. That is 2 million input tokens and 500,000 output tokens in total.

Using the Flash rates above, input costs $1.50 and output costs $1.875. The total is $3.375, or about $3.38. At the listed January 2027 rates, the same token usage would cost $6.75.

This example shows how the calculation works using Standard paid rates. Your actual bill will depend on your requests, and it can include other costs such as caching, grounding, media generation, retries and hosting.

One detail that's easy to miss: a single click in your app can trigger several model calls. A “summarize” button might extract information, write a summary, and then check the result. All of those calls need to be included when you're estimating what the feature will cost.

How much does it cost to publish an AI Studio app?

You may be able to publish without setting up paid billing, but the answer depends on your account and how you deploy the app. You'll also want to consider what happens once people start using it.

Google's Cloud Starter Tier lets eligible accounts publish up to two full-stack apps without setting up a Cloud project or billing account. Eligibility is restricted: Workspace accounts and accounts with current or previous paid Cloud billing can be ineligible, among other reasons.

The standard deployment route requires a Cloud project with billing enabled, and each Studio deployment creates a Cloud Run service. That means the Free Studio plan alone doesn't tell you what your hosting will cost. Google's deployment guide explains the two routes and their requirements.

The rest of your budget depends on what the app does. A simple page that displays information has different needs from a client portal that stores documents, sends emails and generates AI responses.

An easy way to work this out is to walk through the app as a user. When you upload a document, where is it stored? When you ask a question, does the app call Gemini? Each connected service may have its own free allowance and paid pricing, so this walkthrough helps you spot costs before you share the app widely.

Our Google AI Studio tutorial covers the building workflow. Use this pricing guide alongside it when you decide how to publish.

Does Google use free-tier data to improve its models?

Pricing isn't the only difference between free and paid access. Google's data terms also distinguish unpaid services from paid services, with regional exceptions that can affect how your requests are treated.

For unpaid services covered by its general rule, Google can use submitted content and responses to improve products, and human reviewers may process them. The terms tell users not to submit sensitive, confidential or personal information to those unpaid services.

For paid services, Google says it does not use prompts and responses to improve its products. Read the Gemini API terms for the applicable definitions, exceptions and other processing rules.

Sample records are a good choice while you're experimenting. If the app will handle real client data, check the terms for the account and service you're planning to use before connecting it.

Which option should you start with?

For learning and building your first small app, we'd start with Free. You can get a feel for the workflow before committing to a subscription. Our Google AI Studio review goes into more detail about what it's like to build with the tool.

Once you have a project you want to keep working on, the choice becomes easier. A Studio subscription is worth considering if you're regularly running out of access while building. Paid API capacity becomes relevant when your app's requests need it, and hosting costs depend on how you publish and run the app.

You don't have to upgrade everything at once. Start with what your project needs today, then use its actual usage to decide where spending more would help.