Nano Banana 2 — Pro-Grade Images at Flash Speed

Powered by Google Gemini 3.1 Flash Image. Real-world knowledge, up to 14 reference images, dependable in-image text and 4K output — without the wait.

Image Generator
0 / 5000

Advanced Options

Cost 70 creditsRemaining 0 credits
Image Preview
It can look the world up before it paints it
The same face, across a whole series
Translate the poster, not just the copy
One product, as many scenes as you need
What is Nano Banana 2?

Nano Banana 2 is the common name for Google's Gemini 3.1 Flash Image (model ID gemini-3.1-flash-image). It is the workhorse of the Nano Banana family: it pulls in real-world knowledge through Google Search and Image Search grounding, holds a person or a product steady across up to 14 reference images, renders and even translates the text inside a picture, and goes all the way to 4K — while staying on the Flash side of the speed and cost curve. It is the balanced pick for most day-to-day production work.

Four things it does that a plain text-to-image model can't

Nano Banana 2 AI generated photorealistic Shanghai Bund skyline at blue hour with accurate landmark towers

It can look the world up before it paints it

Ask Nano Banana 2 for a real place and you get that place, not a generic impression of it. Grounding through Google Search and Image Search means landmarks, current conditions and factual subjects come from something more reliable than what the model happened to memorise during training. Useful for travel visuals, education content and anything where getting reality wrong is embarrassing.

Prompt
A photorealistic travel photograph of the Shanghai Bund waterfront at blue hour, shot from the promenade on the west bank looking east across the Huangpu River. The skyline is accurate: the Oriental Pearl Tower with its distinct spheres on the left, the bottle-opener Shanghai World Financial Center and the twisting Shanghai Tower behind it, all lit against a deep indigo sky. In the near foreground, the historic stone balustrade of the Bund. Long-exposure boat light trails on the water, wet pavement reflections, crisp architectural detail.
Nano Banana 2 AI generated four-panel storyboard keeping one character consistent across all panels

The same face, across a whole series

Feed Nano Banana 2 up to 14 references — as many as 10 for objects and 4 for people — and it keeps track of who is who and what the product looks like. That is what turns one-off image generation into a series: the same character through four panels, the same bottle across a dozen scenes, without redrawing the brief every time.

Prompt
A four-panel storyboard laid out as a 2x2 grid with thin white gutters between panels. The SAME character appears in all four panels and must stay recognisably identical: a young woman with short copper hair, round tortoiseshell glasses, and a rust-orange canvas jacket. Panel 1: she sits on a train by the window at dawn, sketchbook on her lap. Panel 2: she walks a narrow alley market carrying the sketchbook. Panel 3: she sits on stone steps drawing, warm noon light. Panel 4: she stands on a rooftop at dusk, city lights behind her. Consistent warm cinematic colour grade across all four, identical face and jacket in every panel.
Nano Banana 2 AI generated English and Simplified Chinese versions of the same running shoe poster side by side

Translate the poster, not just the copy

In-image localization is where Nano Banana 2 earns its keep: it takes a finished layout and re-renders the text in another language while the photograph, the colour blocks and the typographic hierarchy stay put. One creative, several markets, without rebuilding the artwork each time. Proofread before it ships — but the structure survives the trip.

Prompt
Two versions of the same printed advertisement poster lying side by side, shot straight down, equal size. Both share an identical design: a pair of white running shoes on a bold coral background, same photograph, same layout, same colour blocks. The LEFT poster is English — a large headline reading exactly "RUN LIGHTER" with a smaller line beneath reading exactly "New Season Collection". The RIGHT poster is Simplified Chinese — the same design with the headline replaced by "轻盈开跑" and the smaller line replaced by "当季新品". The Chinese version keeps the same typographic hierarchy and alignment as the English one.
Nano Banana 2 AI generated e-commerce lifestyle photo of a sage-green insulated bottle on a mountain trail boulder

One product, as many scenes as you need

Upload the product, then let Nano Banana 2 move it anywhere: a trailside boulder, a studio sweep, a kitchen counter. The shape, finish and proportions come along unchanged, so the catalogue still looks like the same object. Cheap and quick enough that testing thirty backdrops is a normal Tuesday rather than a shoot.

Prompt
A commercial e-commerce lifestyle photograph of a matte sage-green insulated water bottle with a bamboo lid, standing on a sun-warmed granite boulder beside a mountain trail. Behind it, soft out-of-focus pine forest and morning haze. The bottle itself is tack sharp and rendered as a real manufactured object: consistent matte texture, a crisp anodised rim, a subtle seam under the lid, believable proportions. Warm directional morning light from the right, natural colour, shallow depth of field.
Key Features
  • Nano Banana 2 runs on Google Gemini 3.1 Flash Image (gemini-3.1-flash-image)
  • Google Search and Image Search grounding for real-world subjects
  • Up to 14 reference images — 10 for objects, 4 for people
  • In-image text rendering and localization across languages
  • 512px, 1K, 2K and 4K output from the same model
  • Conversational multi-turn editing that keeps the earlier context
  • Minimal and High reasoning modes let Nano Banana 2 trade speed against care
Pricing and Credits

Nano Banana 2 sits between Nano Banana and Nano Banana Pro at roughly $0.067 per 1K image. On this site both text-to-image and image-to-image are billed in credits, with the cost shown before every generation. New accounts get free credits on sign up.

What Nano Banana 2 is great at

Campaign concepts at volume

Nano Banana 2 generates ad directions, campaign visuals and social variants fast enough to test many ideas instead of defending the first one.

Product and e-commerce imagery

Keep the product exactly as it is while the background, lighting and setting change around it.

Global campaign localization

Take one finished creative into other languages — Nano Banana 2 re-renders the text inside the image, not just the caption underneath.

Infographics and knowledge visuals

Combine world knowledge and search grounding to turn dense information into something people can actually read.

Characters and serial content

Hold a character's look across locations, outfits and story beats — comics, IP work, ongoing social series.

Image engine inside a product

Low latency and high throughput make Nano Banana 2 practical as the generation layer behind your own app's create-edit-regenerate loop.

How to use Nano Banana 2
  1. 1

    Open the generator

    The generator above is set to Nano Banana 2 already. Sign in to claim your free credits — no payment needed to try it.

  2. 2

    Say what the image is for

    Describe the subject, but also where it will be used — product listing, banner, poster. Naming the job helps it pick composition and text hierarchy.

  3. 3

    Add references if you have them

    Switch to image-to-image and upload the person, product or style you want carried over. Label what each reference is for in the prompt.

  4. 4

    Keep editing in place

    Got something close? Ask for the next change instead of starting over — "same person, move it to a night street", "keep the layout, headline in Chinese".

Nano Banana 2 vs Nano Banana Pro
Nano Banana 2Nano Banana Pro
Best forEveryday production at volumeFlagship commercial assets
SpeedFlash tier, low latencySlower — it plans before drawing
Visual fidelityHighHighest in the family
Reference imagesUp to 14 (10 object / 4 person)Up to 14 (6 object / 5 person / 3 style)
In-image textDependable, multilingualMost precise, best for tight layout
Search groundingWeb + Image SearchWeb Search
Max resolution4K (from 512px)4K (from 1K)
Price per image≈ $0.067≈ $0.134
Nano Banana 2 — Frequently Asked Questions
Common questions about Nano Banana 2 and imagesv2.ai. Have more questions? Contact us at support@imagesv2.ai.

Nano Banana 2 is the common name for Google’s Gemini 3.1 Flash Image (model ID gemini-3.1-flash-image), the Flash-tier workhorse of the Nano Banana family: near-flagship quality at Flash speed and price. It can pull in real-world knowledge through search grounding, holds a subject steady across a stack of reference images, and draws text inside the picture far better than the Flash generation before it. You will also see it written nano banana2, nano-banana-2 or Nano Banana 2 AI — same model. imagesv2.ai runs it in the browser, no Google account or API key needed.

No — different models, released close together. Nano Banana Pro is Gemini 3 Pro Image, the flagship: highest fidelity and the most precise layout, 140 credits per image here. Nano Banana 2 is Gemini 3.1 Flash Image: faster, 70 credits per image, and close enough on most jobs that Pro is worth saving for final deliverables. Both sit in the model picker above.

Free credits land in your account when you sign up on imagesv2.ai — no card, no API key. After that Nano Banana 2 costs 70 credits per 1K image, 105 at 2K and 155 at 4K, with a 20-credit surcharge when you upload an image and edit it. At 1K that puts it between Nano Banana at 40 and Nano Banana Pro at 140, which is exactly where its output sits.

Nano Banana 2 is the faster, cheaper option, and it is unusually strong for its tier at real-world knowledge, multi-reference consistency and text inside the image. GPT Image 2 is OpenAI’s flagship: more exact on long, detailed briefs and on typography-heavy layouts, at a higher price per image. Producing at volume, Nano Banana 2 is the sane default; when a single frame has to be perfect, run it through GPT Image 2 as well. imagesv2.ai supports both from one picker, so comparing costs you a prompt, not a subscription.

The original Nano Banana (Gemini 2.5 Flash Image) is cheaper and a touch faster, but it is noticeably weaker at text inside the image and at holding a subject across many references. Nano Banana 2 adds search grounding, many more reference images, dependable in-image text and translation, and a higher ceiling on detail. Keep the original for high-volume drafts; move up to Nano Banana 2 when the picture has to say something specific.

Google’s model accepts up to 14 in one task; the uploader on imagesv2.ai caps a single generation at 9 images, 5 MB each. Mix them freely — a product, a person, a background, a style plate. As always, a few clean references that agree with each other beat a pile that does not.

Much better than the Flash generation before it — headlines, labels, price tags and short blocks of copy come out readable in Latin and CJK scripts alike, and it can restate the text of an image in another language. Dense small type is still where every image model struggles, so proofread anything you publish.

When search grounding is in play, Nano Banana 2 can lean on current public web information instead of training memory alone — useful for real places, products and public figures. Treat any number, date or claim it draws as a draft and verify it before you publish.

Pick 1K, 2K or 4K in the generator above, alongside the aspect ratio (Auto plus the 14 ratios Google accepts — 1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3, 21:9 and the rest). Credits follow the resolution because Google's price does — 70, 105 and 155 respectively, shown before you press Generate. 2K comes back around 2700px on the long edge and 4K around 5500px. Settling the composition at 1K and re-running only the keeper at 4K is usually the cheapest route to a finished asset.

Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image) is the stripped-down sibling and the cheapest model in the family here at 35 credits per image — made for thumbnails, ad variants and quick drafts where throughput matters more than polish. It is in the same model dropdown above whenever you want to burn through options cheaply.

No. imagesv2.ai calls Gemini 3.1 Flash Image for you — no Google account, no AI Studio project, no API key, no billing setup. Sign in, write the prompt, press Generate. Everything you make stays in your history.

Yes — what you generate here is yours to use commercially: ads, listings, social content, printed material, merch. You need the rights to anything you upload as a reference, including people’s likenesses. Google also embeds an invisible SynthID marker in Gemini output to flag it as AI-generated; it does not change how the image looks.

Images generated on free credits download with a tiled imagesv2.ai watermark. Any purchase — monthly plan, yearly plan or a one-time credit pack — unlocks watermark-free downloads on your account, including for images you generated earlier.
Start generating with Nano Banana 2

Open the generator above, claim your free credits, and try Nano Banana 2 online. Switch to Nano Banana Pro from the same picker whenever a job needs the flagship.