Nano Banana 2

Generate faster, higher-quality images on Nano Banana 2, grounded in live Google Web and Image Search. You describe the image; the agent translates it into the model’s own language, so nothing gets lost on the way.

Trusted by teams at
Phantom X
Wonder Studios
Google
NVIDIA
Salesforce
Visa
IBM
monday.com
Cisco
Mastercard
Meta
LinkedIn
Netflix
Ford
Hilton
Siemens
HubSpot
Phantom X
Wonder Studios
Google
NVIDIA
Salesforce
Visa
IBM
monday.com
Cisco
Mastercard
Meta
LinkedIn
Netflix
Ford
Hilton
Siemens
HubSpot

What sets Nano Banana 2 apart

It looks things up before it draws

Nano Banana 2 pulls live information from Google Web and Image Search while it generates, so real people, places, products, and events render from current references, not a guess from old training data.

Speed that changes how you work

Prompts process fast enough that generating is cheaper than deliberating. Ten directions get explored in the time one used to take, and the deadline stops deciding the creative.

Quality that does not pay for the speed

Detail, color accuracy, and visual fidelity hold at pace. This is the model for work that ships today and still has to look like it was not rushed.

Formats other models do not offer

Beyond the standard ratios, Nano Banana 2 generates at 1:8, 8:1, 1:4, and 4:1: banners, panoramas, skyscrapers, and strips, composed for the shape rather than cropped into it.

Consistent sizing when you need it

512×512 output gives a fixed, uniform canvas for icon sets, avatars, and grids where every image has to match the next one exactly.

It understands the whole prompt

The model reads for meaning, not keywords: a complex brief with context, relationships, and intent lands as one coherent image instead of a collage of nouns.


What Nano Banana 2 is best used for

Infographics and diagrams

Diagrams come out with structure that reads: architectural diagrams, process visuals, data graphics. Fast enough to keep reworking the layout until it explains the thing clearly.

infrographics nano banana 2

Real-world and current visuals

When the image has to show a real product, a real venue, or this week’s event, search grounding makes the reference hunt part of the generation.

Screenshot 2026-07-29 at 3.30.09 PM.png

Social content at volume

This is the volume model for social: images for posts, stories, and covers at the pace a calendar demands.

enhanced-clarity-nano-banana-2-final

Cartoons, comics, and storyboards

The same character, the same style, panel after panel, from single strips to full boards. Describe the grid and the beats, and the page comes back composed.

cartoons-comics-nano-banana-2

Wide, tall, and unusual formats

The ratios most models cannot compose for: site banners, channel art, panoramic headers, vertical posters. The format is generated, not cropped.

Untitled design (4).png


How Nano Banana 2 works with invideo agents

On invideo, Nano Banana 2 sits inside an agentic workflow where the agent routes work to it, writes its prompts, and keeps your look on every generation. Here is how it works in practice:

The agent picks the right model intelligently.

Nano Banana 2 sits on a roster of 200+ models, and the agent routes work to it where it wins: speed, volume, live-world accuracy, and formats nothing else generates.

You say what you mean, the agent says it precisely.

Nano Banana 2 reads whole-prompt meaning, and it rewards specifics: the asset, the exact text, the mood, the ratio. You give it in plain language, and the agent writes the version the model performs best on, and pulls in live Search references where they help.

Inside invideo, generating an image is just the start.

A frame made here can open a UGC ad, seed an explainer, or set the look for a film scene. The agent hands it to Seedance 2.0 as a subject reference or to Veo 3.1 as a first frame, and the still you chose becomes the video you were after.

You always stay in control.

You set how much the agent does on its own: let it generate images freely, or see every prompt before it hits generate. That setting is yours to make, and yours to change.

Helping creatives stay creative

Multiplayer mode

Collaborate in real time with live cursors to show what everyone's working on.

RRebeccajust now
Can we push the train arrival 2 sec later? Feels rushed.
AAiko2m
Love the warmth here — keep this lighting for the reunion shot.

Storyboarding

Turn any script or idea into a shot-by-shot plan, then tweak as needed before generating.

Script writing

Write your script inside invideo, and ask an AI co-writer for help if you'd like.

Timeline editor

Picture Premiere Pro with full AI.

Build your own agents

Create custom agents to fill specific roles like cinematographer, music designer, and more.

From solo creatives to creative enterprises

Private & SecureSOC 2 & ISO AlignedGDPR Compliant

World-class investors stand behind invideo.

Backed by the firms behind Stripe, Spotify, Flipkart, and ByteDance.

Pricing

All paid plans include:

Access to 200+ image, video, audio, music models including Seedance 2.0, Veo 3.1, Kling 3.0, Nano banana pro & Elevenlabs music.

Access to top stock providers like iStock, Storyblocks & more.

Model & agent prices are subject to change.

On-demand credit top-ups available.

Nano Banana 2 FAQs

What is Nano Banana 2?

Nano Banana 2 is Google’s fast image generation model, built for speed and scale while keeping near-flagship quality. It grounds generations in live Google Web and Image Search, understands complex prompts semantically, and supports formats from square to 8:1 panoramic. On invideo, it runs inside the agent’s roster.

What new aspect ratios does Nano Banana 2 support?

1:8, 8:1, 1:4, and 4:1, alongside the standard ratios: ultra-wide banners, tall verticals, and panoramic strips composed natively for the shape.

What output resolutions are available with Nano Banana 2?

Nano Banana 2 now supports 512×512 resolution, providing clear, sharp, and detailed images suitable for both digital and print applications.

How does Google Image Search enhance image generation?

The model references real-world visuals while generating, so a prompt about a real product, place, or recent event draws on current imagery instead of training data alone. Fewer off-brand guesses, less manual reference hunting.

How do I write effective prompts for Nano Banana 2?

Short and specific beats long and vague: the asset type, the subject, the mood, any exact text, and the ratio. On invideo, you skip this: describe the image in plain language and the agent writes the prompt the model reads best.

What are the best use cases for Nano Banana 2?

High-volume and fast-turnaround work: social content, infographics and diagrams, comics and storyboards, real-world visuals that need current accuracy, and wide or tall formats other models cannot compose.

How is Nano Banana 2 different from Nano Banana Pro?

Nano Banana 2 is the speed and scale model; Nano Banana Pro is the quality flagship with the strongest reasoning, text, and 4K output. Most work runs perfectly on 2; Pro earns its cost when precision decides the outcome. On invideo both are on the roster and the agent picks per task.

Is Nano Banana 2 free on invideo?

Nano Banana 2 runs on invideo’s plans, with generations drawing on your plan’s credits.

Do I need to learn Nano Banana 2’s prompting to use it on invideo?

No. The model rewards brief, specific prompts, and the agent writes them for you, grounding and format included. You describe the image the way you would to a person.