Nano Banana 2.1

Nano Banana 2.1 (model code gemini-nano-banana-2.1) is Google's efficient image generation and conversational editing model. It updates Nano Banana 2 (Gemini 3.1 Flash Image) and is a more efficient counterpart to Nano Banana Pro (Gemini 3 Pro Image).

Google's model card says Nano Banana 2.1 maintains Flash-level speed and cost while improving visual quality, prompt adherence, character consistency across multiple editing turns, and text rendering. It also fixes tiling artefacts at 2K and 4K for wide aspect ratios.

Nano Banana 2 remains the lighter, cheaper tier, while Nano Banana Pro is the higher-fidelity option for complex professional work. All three are available in the same generator, so you can try your prompt with each before deciding.

Text-to-image costs 5 credits at 1K, 8 credits at 2K and 10 credits at 4K. Adding a reference image costs 1 extra credit at each resolution, so image-to-image costs 6, 9 or 11 credits. Each generation returns one image. Check the selected model and displayed credit cost before generating.

Choose 1K, 2K or 4K, and save the image as JPG or PNG. The default is 1K.

There are 15 aspect-ratio options: 1:1, 4:5, 5:4, 3:4, 4:3, 2:3, 3:2, 9:16, 16:9 and 21:9; panoramic 4:1 and 1:4; tall 8:1 and 1:8; and auto.

This page accepts up to 10 JPEG, PNG or WebP images, up to 30 MB each. Google's model card lists support for up to 14 references, with character consistency for up to 4 characters and object fidelity for up to 10 objects.

Use sharp, well-lit images and explain what each one is for, such as the product shape, person, pose, material, colour palette or background. State what must not change. A few clearly described references are often more useful than a large, mixed set.

Descriptions can be up to 20,000 characters. A useful order is: subject, scene, framing, lighting, materials, styling, any text to include in the image, constraints and intended use.

Start with what needs to be right, such as the product, face, wording or space for a headline, then add the style. If a detail must stay exactly as it is, say so clearly.

Read any text in the image, check small faces and hands, and confirm product labels, logos and factual details. If anything is wrong, try again with a clearer instruction. Check each result before using it.

This page accepts images and text as input. Google's model card also lists video and PDF input for the Gemini API, but this page's controls do not include Web or Image Search grounding or configurable thinking levels.

To continue, keep the same references and change one thing at a time. Draft at 1K, compare a few directions, then generate the image you want to keep at 2K or 4K. Export as JPG for a smaller file or PNG when you need that format, and check the result after each change. Each run uses credits at the displayed cost for the selected model and resolution.