Fix Gemini Image Generation Refusals: Sign-in, Age, Model

Three things to check first: sign-in, age, model

Image generation in the Gemini app requires signing in.

Google's help page says you must sign in to the Gemini app to use it, and when RS AI DESK asked signed-out gemini.google.com to "draw a single white circle on a blue background" on 24 September 2026, it replied that it could not generate images right now.

If you are signed in and still refused, check age and model settings. Personal accounts must be at least 13 (or the applicable age in your country) to generate images and at least 18 to edit them. With the model set to Flash-Lite, the app uses Nano Banana 2 Lite, which is tuned for speed and described as unsuitable for many reference images or edits.

With Flash or Pro selected, the app uses Nano Banana 2, which accepts several reference images, and the more precise Nano Banana Pro is available for redoing images on a Google AI plan with the model set to Pro.

Refusal causes by situation

Recovery steps

1. Check that you are signed in to gemini.google.com with a Google account; if you see a sign-in screen, sign in first.
2. In the model picker next to the input box, switch from Flash-Lite to Flash or Pro if Flash-Lite is selected.
3. Choose Images in the top-left sidebar, or start the prompt with "draw" or "create".
4. Remove real names, other people's works and anything that may fall under the prohibited-use policy, then describe subject, style and background in detail.
5. Hover over the generated image, confirm that Download full size appears, and save the file.

A model change applies only to requests made after it. Google's help page says to switch the Gemini model before making the image request to use another Nano Banana model, so if you changed the model in an existing chat, send the request again.

Describe the style you want (realistic, watercolor, cartoon and so on), the subject, what it is doing and the background. If the image must contain text, give the exact words in quotation marks.

How to confirm it works

It works when the image appears in the chat and Download full size saves the file. Per Google's help page, Google AI plan subscribers download at 2K and others at 1K, so the file size also tells you whether your plan is applied. Paid subscribers can get a Nano Banana Pro version from More at the bottom right of the image, then Redo with Pro.

To change an image you already made, open it from Library in the left sidebar, click Chat below it and type the change. You can also upload your own image and ask for edits, but editing is available only to users aged 18 or older.

When it still fails

If the same request keeps being refused, collect the account type (personal, school or work), model setting, plan, full prompt and screenshots and contact Gemini support. For school or work accounts, first ask the admin whether Gemini image features are allowed, and if they are blocked use another tool your organization's policy permits.

Availability also depends on the app's supported languages and countries.

Conditions that are easy to miss

Once the daily Nano Banana 2 image limit is reached you cannot redo images with Nano Banana Pro either, so repeating Redo with Pro after the limit produces nothing. Users under 18 on school accounts cannot use Export to Docs.

Before publishing or using generated images at work, check the prohibited-use policy and that you are not infringing others' copyright or privacy rights.