Nano Banana is Google's family of Gemini image models: the original, Nano Banana Pro, Nano Banana 2 and Nano Banana 2 Lite. Each one generates and edits images from plain-English instructions. Pro, 2 and 2 Lite run in the generator below; Google's Gemini API retires the original on October 2, 2026. New accounts get 20 free credits, and a 1K image on 2 Lite costs 5.
GALLERY
CAPABILITIES
Feature 01
Google launched the first model, Gemini 2.5 Flash Image, on August 26, 2025, and its nickname stuck to the whole line. Every member takes a text prompt, one or more images, or both, and returns a finished picture. Because the models are built on Gemini, they bring its world knowledge along: ask for a botanical plate of a monstera or a street in 1920s Paris and the details come out plausible without a paragraph of explanation.
Feature 02
Keeping a person or a mascot recognisable across images was the headline feature at launch, and it is still the reason most people try these models. Put your character in a café, then on a mountain, then in a comic panel, and the face, hair and outfit carry over. Nano Banana Pro goes furthest: Google says it keeps the resemblance of up to five people in one composition.
Feature 03
No masks, no layers. Type "blur the background", "remove the stain on the shirt" or "make it golden hour" and Nano Banana changes only that. You can keep going in the same thread: fix the lighting, then swap the jacket, then crop to 9:16. Prompts work in many languages, and the newer models can also translate the text already inside an image.
Feature 04
Upload a product, a model and a location and ask for one photo of all three together. The original Nano Banana accepts up to 5 references per request on Ricebowl; the newer three take up to 8 here. Google's API goes further for the Gemini 3 models, with up to 14 reference images, but the Ricebowl uploader is capped at 8 today.
Feature 05
Three of the four are Flash models, designed for low latency so you can try ten ideas in the time a heavier model takes for two. Nano Banana 2 Lite is Google's fastest and cheapest image model; one developer quoted on Google DeepMind's page measured it at about 2.7 times faster than 2 for 1K images. Pro is the exception: it runs a thinking pass on every request, which costs time and buys accuracy.
PROMPT EXAMPLES
Prompt
Keep the person exactly as in the photo. Change the background to a sunlit Mediterranean terrace with white walls and bougainvillea, soft afternoon light, same pose and outfit, 9:16.
Reference
Result
Prompt
Restore this old photo: remove scratches and dust, fix the faded colours, sharpen the faces slightly, keep the original clothing and composition, 1:1.
Reference
Result
USE CASES
Drop a phone photo of the product, describe the new setting, and get a clean catalogue shot. Do the whole range at 1K with Nano Banana 2 Lite for 5 credits each, then redo the hero image at 2K with Nano Banana Pro.
Upload a selfie and ask for a studio headshot, a LinkedIn photo or a stylised avatar. Character consistency keeps it recognisably you.
Remove scratches, repair faded colour and sharpen faces on scanned prints. Say what to keep, such as the clothing and the composition, and the model leaves it alone.
Keep one character consistent across a dozen panels or storyboard frames, changing only the scene and the pose.
Pro and 2 can render and translate text inside an image, so one poster becomes versions for several markets without rebuilding the layout.
WHAT USERS SAY
01
Most people arrive for the editing, not the generation: describe a change and it happens, with no selection tools. That is what made the first release go viral.
02
Four models share one nickname and users regularly compare the wrong ones. The short version: Pro for quality, 2 for speed with search, 2 Lite for volume, and the original is on its way out.
03
Character consistency is the most praised capability, especially for headshots and storyboards.
04
The original struggled with dense lettering. Pro fixed most of that, which is why text-heavy work moved there.
WHICH ONE
Nano Banana Pro for any final image and anything with text: it is Google's premium model and on Ricebowl it costs 8 credits at 2K, less than 2 (12). Use 2 when you want Flash speed or Google Image Search grounding. Use 2 Lite for bulk 1K drafts at 5 credits. Skip the original: Google's Gemini API retires it on October 2, 2026.
Final images and anything with text. 8 credits at 1K or 2K and 14 at 4K on Ricebowl, cheaper than 2 at the larger sizes.
Flash speed plus Google Search and Image Search grounding. 8, 12 or 18 credits for 1K, 2K or 4K.
Bulk 1K drafts at 5 credits each. Google's recommended replacement for the original.
Retiring. Google's Gemini API shuts down Gemini 2.5 Flash Image on October 2, 2026; Vertex AI follows on March 15, 2027. Move existing prompts to 2 Lite.
WHAT SETS IT APART
01
You do not need a Google account, a Gemini subscription or an API key. In the Gemini app the default is now Nano Banana 2, and only paid Google AI Pro and Ultra subscribers can reach Pro. On Ricebowl you pick Pro, 2 or 2 Lite from the model menu and pay per image in credits.
02
Switch models without re-uploading your references and compare results side by side. Credits are shown before you press generate, so the difference between a 5-credit draft and a 14-credit 4K render is always visible.
OUR TAKE
Default to Nano Banana Pro. On Ricebowl it costs the same 8 credits as 2 at 1K and less at 2K and 4K, and Google ranks it top for quality and text. Drop to 2 Lite only for bulk 1K drafts, and to 2 only when speed or image-search grounding matters more than the last bit of quality.
No. Google has deprecated Gemini 2.5 Flash Image, and the Gemini API shuts it down on October 2, 2026, telling developers to move to Nano Banana 2 Lite for better quality, faster generation and lower API pricing ($0.0336 vs $0.039 per image). On Ricebowl both cost 5 credits, so switch to 2 Lite in the model menu now.
Neither. On Ricebowl a Google image costs the same whether you start from a prompt or a photo, and reference images add nothing. The saving comes from the model and size: 5 credits for a 1K image on Nano Banana 2 Lite, 8 for a 2K image on Nano Banana Pro.
COMPARISON
| Feature | Nano Banana family (Pro / 2 / 2 Lite) | GPT Image 2 |
|---|---|---|
| Developer | Google DeepMind | OpenAI |
| Keeping a person consistent | Core strength; Pro holds up to 5 people | Strong on a single reference |
| Text inside images | Pro is Google's best for legible text | Much improved; OpenAI notes placement can slip |
| Editing style | Plain-English edits, relighting, translation | High-fidelity reference edits, mask as guidance |
| Search grounding | Web search; 2 adds image search | None |
| Largest output | 4K on Pro and 2; 1K on 2 Lite | 4K |
| Aspect ratios on Ricebowl | Auto, 1:1, 16:9, 9:16 | 11 ratios from 21:9 to 9:21 |
| Ricebowl credits, 1K / 2K / 4K | Pro 8 / 8 / 14; 2 8 / 12 / 18; 2 Lite 5 (1K) | 6 / 10 / 16 |
| Pick it for | Portraits, characters, text-heavy posters | Product shots from a reference, wide banners, single-line titles and labels |
Both families run on Ricebowl; switch in the model menu. Credits are per image; GPT Image 2 adds 1 credit for each reference after the first, Google's models do not.
Try Nano Banana Free01
Open the model menu in the generator above and choose Nano Banana Pro, Nano Banana 2 or Nano Banana 2 Lite. Create a free account: it comes with 20 credits and a 1K image on 2 Lite costs 5, with no card.
02
Add a photo to edit, or several to blend, then write what should change or what to create. Keep instructions in plain sentences.
03
Press generate, review the result, and send a follow-up instruction to adjust it. Download the final image from the card.
REVIEWS
Notes from people who use Google's image models for everyday edits.
Clara JensenI photograph my ceramics on the kitchen table and turn them into studio shots in one sentence each.
Etsy seller
Rahul MehtaMy comic character looks the same in every panel. That used to take me hours of redraws.
Freelance illustrator
Beatriz SouzaDrafts on Nano Banana 2 Lite, finals on Nano Banana Pro. The credit difference is right there before I press generate.
Social media manager
Tom FischerI restored forty scanned family photos in an evening. Faces stayed faces.
Hobby genealogist
Yuki SatoHalf my candidates now send headshots made from a selfie. I made mine here too.
Recruiter
One balance, every model. Credits work across all 19 video models — no watermark, no tier-locked models, commercial use on every paid plan.
Save 30% billed annually
Pay yearly and save more.
Enough credits to try every model and finish a few clips each month.
$19.9/mo$29.9
Billed yearly at $238.8. Renews yearly, cancel anytime.
Save $120/yr vs monthly
12,000 credits
≈ 38–600 videos, depending on the model
≈ 857–4000 images, depending on the model
1.99¢ / credit
2 parallel tasks
Steady output for solo creators going pro — every model, a comfortable budget.
$34.9/mo$49.9
Billed yearly at $418.8. Renews yearly, cancel anytime.
Save $180/yr vs monthly
24,000 credits
≈ 76–1200 videos, depending on the model
≈ 1714–8000 images, depending on the model
1.75¢ / credit
3 parallel tasks
High-volume production for prolific creators and small teams.
$62.9/mo$89.9
Billed yearly at $754.8. Renews yearly, cancel anytime.
Save $324/yr vs monthly
60,000 credits
≈ 190–3000 videos, depending on the model
≈ 4285–20000 images, depending on the model
1.26¢ / credit
5 parallel tasks
Flexible monthly subscription.
Enough credits to try every model and finish a few clips each month.
$29.9/mo
Renews monthly, cancel anytime.
1,000 credits
≈ 3–50 videos, depending on the model
≈ 71–333 images, depending on the model
2.99¢ / credit
2 parallel tasks
Steady output for solo creators going pro — every model, a comfortable budget.
$49.9/mo
Renews monthly, cancel anytime.
2,000 credits
≈ 6–100 videos, depending on the model
≈ 142–666 images, depending on the model
2.50¢ / credit
3 parallel tasks
High-volume production for prolific creators and small teams.
$89.9/mo
Renews monthly, cancel anytime.
5,000 credits
≈ 15–250 videos, depending on the model
≈ 357–1666 images, depending on the model
1.80¢ / credit
5 parallel tasks
Top up credits valid for 12 months.
One-time top-up — perfect for occasional needs.
$49.9
One-time purchase, valid 12 months.
1,000 credits
≈ 3–50 videos, depending on the model
≈ 71–333 images, depending on the model
4.99¢ / credit
Larger top-up with 20% bonus — best per-credit value for one-time buyers.
$198
One-time purchase, valid 12 months.
4,800 credits
≈ 15–240 videos, depending on the model
≈ 342–1600 images, depending on the model
4.13¢ / credit
5 parallel tasks
What the four models are, what they cost on Ricebowl, and which one to pick.
Try Nano Banana FreeYes, to start. New Ricebowl accounts get 20 free credits on sign-up, and a 1K image on 2 Lite costs 5 credits (the original is the same price until Google's Gemini API retires it on October 2, 2026), so your first images need no card. After that you pay per image with credits from a plan or a top-up.
It is the nickname for Google's Gemini image models. There are four: the original (Gemini 2.5 Flash Image, August 2025), Pro (Gemini 3 Pro Image, November 2025), 2 (Gemini 3.1 Flash Image, February 2026) and 2 Lite (Gemini 3.1 Flash Lite Image). All generate and edit images from text and image inputs.
It made plain-English photo editing, character consistency and multi-image blending mainstream. Today there is little reason to pick the original Nano Banana itself: Google's Gemini API shuts it down on October 2, 2026 (Vertex AI keeps it until March 15, 2027) and Google recommends 2 Lite, which costs the same 5 credits on Ricebowl.
Quick 1K edits: background swaps, object removal, simple restorations and avatar variations. For new projects, run the same jobs on 2 Lite, its designated replacement, or on Pro when quality and text matter.
Photorealism, studio product shots, watercolour, oil painting, anime, comic, pixel art, 3D render, claymation and more. Name the style in the prompt, or upload a reference image and ask the model to match it.
It already is. Nano Banana 2 runs in the same generator at 8, 12 or 18 credits for 1K, 2K or 4K. See the dedicated Nano Banana 2 page for its features and how it compares with Pro.
Yes, through Google's Gemini API and Vertex AI. Model IDs are gemini-3-pro-image, gemini-3.1-flash-image, gemini-3.1-flash-lite-image and gemini-2.5-flash-image. Google's per-image prices range from $0.0336 for 2 Lite at 1K to $0.24 for Pro at 4K.
Nano Banana Pro for final images and anything with text; on Ricebowl it is also cheaper than 2 at 2K and 4K. Nano Banana 2 for fast iteration with search grounding. Nano Banana 2 Lite for large batches of 1K drafts. Not the original, which Google's Gemini API retires on October 2, 2026.
Yes. Upload a clear photo and describe the change, such as a new background, outfit or lighting. Nano Banana keeps faces recognisable. Only upload photos of yourself or people who have agreed, and note that every output carries Google's invisible SynthID watermark.
Common image files work as references. On Ricebowl you can attach up to 5 references per request for the original Nano Banana and up to 8 for the other three.
Right here on Ricebowl. Pro, 2 and 2 Lite run in the generator at the top of this page; Google's Gemini API retires the original on October 2, 2026. Sign-up gives you 20 free credits, enough for four 1K images on 2 Lite, and the same workspace also runs GPT Image 2 and 30+ other image and video models.
RELATED
MORE MODELS
Google's Nano Banana image models, plain-English editing, and the credit cost shown before you generate.
Try Nano Banana FreeFREE CREDITS ON SIGN-UP · PRO / 2 / 2 LITE · UP TO 8 REFERENCES · NO GEMINI APP
Try Nano Banana on Ricebowl
Try Nano Banana Free