Make a blog featured image in Gemini and download it at 2K

Image limits by account type

What you can generate, and the resolution you can download, depends on your Google account type and plan. Check the limits for your tier before you make an image for a work or commercial blog.

Users under 18 can generate images, but partial edits and re-editing a generated image are restricted. On a Workspace account for work, an organization administrator has to enable access to the Gemini app before you can open it.

A prompt for a blog featured image

Getting clean on-image text and a usable graphic out of a generative model takes a specific prompt structure. When the result is not what you wanted, copy the example below and change only the graphic elements and the wording.

Create a featured image for an IT blog in a watercolour style. Place a robot wearing glasses working at a laptop in the centre of the image. The background should be a clean office in a light blue and white gradient. Render the words "AI 가이드" in Korean across the top of the image, spelled correctly and clearly legible.

Long wording on an image is where text rendering breaks. Keep the headline in the prompt to one or two clear words.

Generating and downloading a thumbnail in five steps

This is the sequence for generating a blog thumbnail on the official Gemini site in a desktop browser and saving the original file.

1. Open gemini.google.com in a desktop browser and sign in with your Google account.
2. Open the sidebar from the top left, then pick a model flagged Nano Banana 2 or Pro in the model selector.
3. In the input box at the bottom, enter a prompt that specifies the image style, the main subject, the background and any text to include, then submit it.
4. When the generated image appears, use the chat box under it to make partial edits, for example "make the background brighter".
5. Hover over the finished image and click Download original size, or open the more menu and choose Export to Docs.

On a paid Google AI plan, the Re-run on Pro button at the bottom right regenerates the image at 2K with finer detail and better text rendering.

Review checklist and policy

Check the spelling of any on-image text and your compliance with Google policy before you upload the result to a blog. An image that fails these checks can have its reach limited or draw a policy enforcement.

• Compare the spelling of the text inside the image against what you asked for.
• Look for distorted detail such as faces and fingers.
• Check the image for anything that infringes someone else's copyright or trademark under Google's Generative AI Prohibited Use Policy.
• Confirm that no real person's likeness or personal information appears without their consent.

Under that policy, requests to synthesize a public figure for impersonation, to produce sexual content, or to produce violent content are blocked automatically.

Diagnosing a failed generation

When a generation request is refused or a refusal message appears, check the wording of your prompt and your account permissions first.

A banned keyword or the name of a copyrighted character in the prompt produces a policy block message. Remove proper nouns and brand names, describe the subject in ordinary words, and try again.

Once you reach the daily Nano Banana 2 generation limit, further generation through Nano Banana Pro is blocked as well. Wait for the limit to reset, or move to an external editor such as Canva or Adobe Express to finish the featured image.

If access is blocked because you signed in with a school or company account, ask your Workspace administrator to approve Gemini access in the admin console.