Fast creative iteration
Develop several directions before committing to a final layout or high-resolution render.
Generate and edit images with Google’s Gemini 3.1 Flash Image model, designed to balance image quality, instruction following, speed, and cost for everyday creative work.
Nano Banana 2 is selected by default, and you can switch to another available image model from the model menu.
Generated content will appear here
Facts reviewed:
Nano Banana 2 is Google’s name for Gemini 3.1 Flash Image. Google describes it as the general-purpose model in the Nano Banana family, balancing performance, intelligence, latency, and cost for image generation and advanced editing.
Google’s current documentation lists 0.5K, 1K, 2K, and 4K output, expanded aspect ratios, improved multilingual text rendering, multiple-reference workflows, configurable thinking, and text-and-image Search grounding. Some provider capabilities are not necessarily exposed in every third-party interface.
The workspace on this page uses the Nano Banana 2 option configured by this platform. Its aspect ratios, reference limit, point estimate, and task availability are platform settings, while the model identity and general capabilities come from Google’s documentation.
A general-purpose choice for rapid generation, editing, and repeatable production workflows.
Develop several directions before committing to a final layout or high-resolution render.
Use layered prompts for subject, camera, environment, style, text, and must-keep constraints.
Upload references to restyle, recompose, localize, or adapt an existing visual.
The platform offers square, portrait, landscape, ultrawide, and extreme-banner formats for this model.
Select the platform output size that fits drafting, publishing, or detailed inspection.
Combine visual direction from several uploads while checking identity, text, and small details after generation.
Draft posters, cards, interfaces, and campaign graphics with translated in-image copy, then proofread every character.
Google documents text and image Search grounding for real-world context; access can depend on the integration and task.
A general-purpose choice for rapid generation, editing, and repeatable production workflows.
Create multiple compositions, colors, or channel formats from one campaign direction.
Test backgrounds and presentation styles while preserving the supplied product as closely as possible.
Draft visuals for different languages and markets, followed by manual copy and cultural review.
Move quickly through storyboards, environments, props, and visual metaphors before final production.
Develop recurring people, pets, or illustrated characters across scenes with explicit continuity instructions.
Explore vertical stories, social posts, presentation backdrops, and extreme banner layouts using supported ratios.
These examples emphasize repeatable direction for everyday production.
Build a repeatable layout with clear variable elements.
Create a set-ready 1:1 social ad for a citrus sparkling water. Bright midday light, cobalt-blue table, one sliced grapefruit, condensation on the can, bold empty headline area at top, energetic but uncluttered commercial photography.
Control visual metaphor and reading order.
Editorial illustration about remote teamwork across time zones: three desks connected by a continuous ribbon of light, morning, afternoon, and night windows, modern geometric style, limited navy and coral palette, horizontal 16:9 composition, no text.
Upload a product image before applying this instruction.
Preserve the uploaded shoe exactly. Replace only the background with a clean pale-gray studio sweep, add a soft natural contact shadow, keep the camera angle and product colors unchanged, and remove all unrelated props.
A fast set of related assets with one reusable direction.
Create three coordinated spring social visuals for a neighborhood coffee shop. Keep the same coral cup, pale mint background, soft paper-cut flowers, friendly daylight, and clear headline space. Vary only the crop and flower arrangement; no visible brand logo.
A composition built for an extreme horizontal placement.
Create an ultra-wide 4:1 launch banner for a compact electric scooter. Place the scooter on the right third, rainy neon city reflections behind it, cyan and lime accents, generous dark negative space on the left for copy, realistic product proportions, no text.
Separate the subject, composition, visual style, required details, and exclusions.
Add only images that clearly communicate what should be preserved or changed.
Choose an available aspect ratio and 1K, 2K, or 4K output, then check the point estimate.
Review text, object count, hands, faces, product geometry, and factual details before reuse.
Google’s API supports a 0.5K tier and configurable thinking, but this workspace intentionally offers 1K, 2K, and 4K only. This page documents the controls available in this platform’s current integration and does not imply identical pricing, latency, quotas, Search grounding, or feature access.
Fast iteration does not remove the need for review. Generated text, people, products, maps, charts, and referenced identities can drift or contain errors. Confirm usage rights for uploads and independently verify any factual visual.
Google states that generated images include SynthID. Model safeguards and provenance signals do not decide whether an upload or output is lawful for your intended use, so rights and publication review remain the user’s responsibility.
Model identity and capabilities were checked against Google’s current documentation and announcement; platform controls come from the live workspace.
Use Google’s Pro image model for more complex professional asset work.
02Compare a separate verified image-generation and editing model.
03Compare image models and choose a workflow from one tool page.
04Improve prompt structure before generating a new image.
Return to the workspace with Nano Banana 2 selected by default, then review every result before publishing.