Yes, there is a free way in. Google says so on the page that announced the model, and the sentence is worth reading in full because almost nobody quotes it correctly.
“Our free-tier users will receive limited free quotas, after which they will revert to the original Nano Banana model. Google AI Plus, Pro and Ultra subscribers receive higher quotas.”
That is the entire official statement about free access. It confirms three things and refuses a fourth. It confirms free quotas exist, that they are limited, and that running out moves you down to the earlier Nano Banana rather than to a paywall. It does not say how many images you get.
What “limited free quotas” covers today

Google’s announcement describes the model rolling out globally in the Gemini app when you pick “Create images” with the “Thinking” model. Subscribers of the three Google AI plans get higher quotas — again, with no figure attached.
What the free tier does not change
Two properties survive whether you pay or not, and both come from the official model page. Free access limits how much you can do, not what the model is.
- Resolution goes up to 1K, 2K or 4K.
- Consistency holds for up to five characters, and fidelity for up to fourteen objects in a single workflow.
Path one: the Gemini app free tier
This is the route almost everyone means by “free Nano Banana Pro”. You open the Gemini app, choose to create an image, and select the Thinking model. That selection is the part people miss; the older model is still a valid choice and produces a different result.
The official announcement also names three narrower doors for the same model, and all three are subscriber routes rather than free ones:
- AI Mode in Search carries it for Google AI Pro and Ultra subscribers in the United States.
- NotebookLM carries it for subscribers globally.
- Flow carries it for Google AI Ultra subscribers.
One honest gap: Google has never published the free tier’s image count. Anything that sounds like a precise daily number is somebody’s observation or somebody’s marketing, not a figure from Google.
Path two: the developer surfaces are a paid preview
The announcement groups the Gemini API, Google AI Studio and Vertex AI under one sentence about starting to roll out. The developer-facing write-up is more explicit, calling it a paid preview.
So the answer for developers is the opposite of the answer for app users. If you are building on it, assume billing. Where the free tier is a quota, the API is a rate. The API guide covers the request shape and what the official pages do and do not disclose about cost.
Worked example: checking a “free Nano Banana Pro” claim
Sites that resell the model usually publish generous-looking free numbers, and their own pricing sits one click away. Here is the check to run before you sign up for one.
1. Does the page name Google's own model, or does it only say "Nano Banana Pro"?
2. Does it link to any Google page at all, or only to its own signup?
3. Does the free number come with a source, or is it stated as fact?
4. Is the free allowance one-off or resetting? The page will say if you look.
5. If you upload a photo, where does it go? Google's pages say nothing about third-party handling.
Case two is the mirror image. A page that names the model “Gemini 3 Pro Image”, links to the official announcement and says the free quota is “limited, official figure not published” has told you more with a smaller number than a page promising thousands of images. Specificity you can verify is the signal.
That is the whole test. Subtract the branding and check what is left.
The watermark question
This is the practical catch nobody puts in a headline. Google states that all media from its tools carry an imperceptible SynthID watermark, and then adds a second layer that depends on your plan.
“In addition to SynthID, we will maintain a visible watermark (the Gemini sparkle) on images generated by free and Google AI Pro tier users, to make images even more easy to detect as Google AI-generated.”
Free output carries a visible mark. So does Google AI Pro output. Google says it removes the visible watermark for Google AI Ultra subscribers and inside the Google AI Studio developer tool. SynthID is not removed at any tier.
How to check an image yourself
Google documents a verification route that does not need any third-party tool. Upload an image into the Gemini app and ask whether it was generated by Google AI, which works through SynthID; Google notes it starts with English-language prompts for images.
Why every free number online is different
Search the question and you will find daily allowances of roughly two, two to four and twenty images for the app, and daily request ceilings of fifty, five hundred and fifteen hundred for AI Studio. Those figures cannot all be right, and none of them traces back to a Google page.
The spread has three causes. Quotas change and blog posts do not. Some numbers describe the older model. Others describe a specific account, region or test window. When a page gives a precise number and no source, you are reading an estimate that has been copied until it looks like a fact.
Frequently asked questions
Is Nano Banana Pro free? Partly. Google confirms free-tier users receive limited free quotas in the Gemini app. The size of that quota is not published.
Is it free forever? Google has published no end date, and also no promise. Read the term as “current policy”, the same way you would read the quota itself.
Which free route should I pick first? The Gemini app, because it is the surface Google describes as shipping the model to free users. From there, the getting-started guide covers a first image end to end.
Can I use free output commercially? Google’s announcement does not address commercial use of free-tier output. Treat that as unpublished rather than assumed, and read the visible-watermark rule as part of the answer.
Do I need a credit card? Nothing in the official material describes a card requirement for the free tier. Third-party sites that ask for one are describing their own signup, not Google’s.
Where do I read the official wording myself? The announcement post is the primary source, and the topic hub collects the model and prompting pages in one place. If a claim about free access is not on one of those pages, its source is somebody else.
The short version: free access is real, limited and undocumented in size. Plan around the direction Google states — quotas, then a fallback to the earlier model — and treat every specific number you read elsewhere as an observation rather than a specification.