Skip to main content
TikTomato
Sign in

Nano Banana 2.1 image generator

Reference images (optional) · Up to 4 images, JPG / PNG / WebP, 5 MB each

Google's newest Nano Banana model, released in October 2026 as an update to Nano Banana 2, with clearer text, more faithful edits from reference images and output up to 4K.

On this page you can add up to 4 reference images (the model itself accepts more). Nano Banana 2.1 has no transparent-background option.

About Nano Banana 2.1

Nano Banana 2.1 is an image generation and conversational editing model from Google, part of the Gemini 3 series. Google released it in October 2026 as an update to Nano Banana 2, keeping that model’s speed and adding improvements in visual quality, prompt adherence, character consistency and text rendering.

It makes images from a text prompt or edits and combines the reference images you give it, at 1K, 2K or 4K. In Google's own side-by-side human evaluations, the model card reports higher scores than Nano Banana 2 for overall preference and for infographic design.

What Google says is new

From the Gemini API model page and the Nano Banana 2.1 model card.

Clearer text and infographics

Google lists enhanced text rendering and more accurate infographic layouts. Posters, labels and diagrams with short text come out well.

Edits that keep your subject

Better character consistency when you combine or edit reference images: the same person or object can move into a new scene.

Quality at 1K, 2K and 4K

Improved visual quality and realism at every output size. Choose 4K when you need fine detail or a large print.

Images generated with Nano Banana 2.1 on this page

Each of these was generated with Nano Banana 2.1 at 1K. The caption is the exact prompt sent to the model.

Food magazine cover titled FLAVOR with a chef plating a dish in a busy kitchen, generated with Nano Banana 2.1
“A glossy food magazine cover. The masthead across the top reads "FLAVOR" in huge bold serif letters. Cover photo: a confident middle-aged Black woman chef with short gray hair and a white chef jacket, plating a dish in a busy restaurant kitchen, steam rising, warm lights and copper pans behind her, natural skin texture, shallow depth of field, photorealistic editorial photography. Three short cover lines in large clean sans-serif type down the left side: "AUTUMN HARVEST", "30-MINUTE DINNERS", "INSIDE THE MARKETS". No barcode, no other text.”
Infographic poster titled How a Tomato Grows with five labeled stages, generated with Nano Banana 2.1
“A clean modern infographic poster titled "HOW A TOMATO GROWS". Five stages from left to right along a gently curved path: seed, sprout, flowering plant, green tomato, ripe red tomato. Each stage has a simple flat illustration and a one-word label underneath in bold sans-serif type: SEED, SPROUT, FLOWER, GREEN, RIPE. Cream background, tomato red and leaf green accents, generous spacing, no other text.”
Elderly woman wrapping yellow tulips at her flower cart on a rainy street at dusk, generated with Nano Banana 2.1
“Candid street photograph of an elderly woman selling flowers from a wooden bicycle cart on a narrow cobblestone street at dusk, light rain, warm shop windows reflecting on the wet stones, she is smiling while wrapping a bouquet of yellow tulips in brown paper. Natural skin texture, 35mm lens, shallow depth of field, documentary style.”
Jar of strawberry jam with a handwritten label reading Strawberry Jam June 2026, generated with Nano Banana 2.1
“A glass jar of homemade strawberry jam on a linen tablecloth next to fresh strawberries and a silver spoon. The jar has a kraft paper label with neat handwritten ink lettering that reads "Strawberry Jam" and below it "June 2026". Soft morning window light, shallow depth of field, photorealistic product photo.”
Four-panel comic of a gray cat chasing a laser dot with speech bubbles, generated with Nano Banana 2.1
“A four-panel comic strip in a 2x2 grid about a gray cat trying to catch a red laser dot. Panel 1: the cat spots the dot, speech bubble "WHAT IS THAT?". Panel 2: the cat pounces and misses. Panel 3: the dot climbs the wall, the cat stares, speech bubble "CHEATER." Panel 4: the cat sleeps on the laser pointer, speech bubble "MINE NOW." Clean ink lines, flat pastel colors, clear hand-lettered speech bubbles.”
Vintage travel poster titled Lake Solvik featuring the woman from the reference photo, generated with Nano Banana 2.1

Reference selfie of a woman with long dark hair, used for the travel poster editReference photo

“Place the person from the reference photo on a vintage 1950s travel poster for a mountain lake town. Keep their face, hairstyle, skin tone and expression recognizable. Flat screen-printed colors, bold simple shapes, the person in the foreground holding a canoe paddle, the lake and pine-covered mountains behind. Large title text at the top reads "LAKE SOLVIK". No other text.”
Panoramic 21:9 view of misty rice terraces at sunrise, generated with Nano Banana 2.1
“Ultra-wide panoramic landscape of terraced rice fields curving around green hills at sunrise, low mist in the valleys, a small wooden farmhouse with a thin line of smoke, a few farmers in conical hats walking along the terrace edges, soft golden light, photorealistic, continuous seamless scene from edge to edge.”

How to prompt Nano Banana 2.1

What follows from the model card and from the examples on this page.

Keep text short and large

Put the exact words in quotes. The model card notes that small text can blur at 1K and long paragraphs are weak, so use a few large words, or choose 2K or 4K for smaller lettering.

Say what to keep when you edit

With a reference image, name what must stay the same (face, hairstyle, the product label) and what should change. Consistency is good but not perfect, so generate a couple of versions.

Describe the whole scene

Subject, setting, light, camera or art style. The model follows detailed prompts closely, so the detail you give is the detail you get.

Pick the shape first

Choose 21:9 for panoramas and 9:16 for phone wallpapers before you generate, rather than cropping afterwards.

Nano Banana 2.1 FAQ

Is this the real Nano Banana 2.1?

Yes. Every image on this page is generated with Google's Nano Banana 2.1, not another model. The examples above were made here with the prompts shown.

How much does an image cost?

Each image uses credits, and the price depends on the output size you choose. The cost is shown on the button before you generate, and new accounts receive free credits.

Can I edit my own photos?

Yes. Add up to 4 reference images and describe what to change or how to combine them. Google says the model supports more reference images; this page accepts 4.

Why is small text sometimes blurry?

Google's model card lists small text at 1K and long paragraphs as known weaknesses. Use fewer, larger words, or switch to 2K or 4K when the text has to be small.

Can I make transparent PNGs?

Not with Nano Banana 2.1: it has no transparent-background option. Use our background remover on the result, or the standard image generator, which can make transparent images.

Nano Banana 2.1 or GPT Image 2.5?

Both are strong general-purpose image models from different companies, and each has its own style. For an important image, try the same prompt on both pages and keep the one you like.

What if my request is declined?

Prompts, reference images and results are checked for sexual and other sensitive content, and flagged requests are declined. Rephrase the prompt and try again. Credits for declined or failed requests are refunded.