Standard
24
credits / image
= $0.024 USD / image
Official: $0.039−38%
Available via API
Nano Banana
Fast, efficient image generation and conversational editing for high-volume creative workflows.
Image Generation
Model Guide
About Nano Banana
Fast, efficient image generation and conversational editing for high-volume creative workflows.
Why choose Nano Banana?
Nano Banana is Google’s name for Gemini 2.5 Flash Image. It is the speed-and-efficiency option in the Gemini image family: a practical fit for high-volume generation, rapid visual iteration, and conversational edits where turnaround matters more than maximum output resolution.
| Core strength | What it means in practice |
|---|---|
| Fast visual iteration | Generate a direction, revise it with plain-language instructions, and compare alternatives quickly |
| Native multimodal understanding | Use text and reference images together instead of translating every visual detail into a prompt |
| Conversational editing | Ask for focused changes while keeping the main subject, composition, or style recognizable |
| Flexible composition | Choose from the Wizzx aspect-ratio controls for square, portrait, landscape, and ultrawide work |
Best-fit workflows
- Social and campaign variations that need many compositions from one creative direction
- Product and lifestyle concepts where a reference image should guide the subject or visual language
- Iterative image editing such as changing a background, color direction, material, or scene context
- Early-stage visual exploration before committing to a higher-resolution production model
Text-to-image or reference-guided editing?
| Start with text | Add reference images |
|---|---|
| You are exploring a new visual direction | Identity, product shape, palette, or composition should remain recognizable |
| Speed and variety matter most | The edit needs visual context that is difficult to describe precisely |
| The prompt can define the whole scene | You want to revise an existing asset rather than recreate it from scratch |
The Wizzx form accepts up to 10 reference-image URLs. Google’s guidance says Gemini 2.5 Flash Image works best with up to three input images, so begin with one strong reference and add more only when each image has a clear role.
Prompting playbook
- State the task first. Say whether you want a new image or an edit.
- Name what must stay stable. For edits, identify the subject, product, pose, layout, or style that should not change.
- Describe the requested change. Be specific about background, lighting, color, material, camera distance, or aspect ratio.
- Iterate one decision at a time. Small follow-up edits are easier to evaluate than a long list of competing changes.
Generation example: Create a clean editorial product photograph of a compact espresso machine on pale travertine, three-quarter view, soft morning window light, restrained warm-gray palette, generous negative space on the left.
Editing example: Keep the product shape, camera angle, and reflections unchanged. Replace the kitchen background with a quiet boutique-hotel breakfast bar at sunrise and shift the accent color from red to deep forest green.
Current Wizzx controls
| Control | Use it for |
|---|---|
| Prompt | The generation brief or the exact edit instruction |
| Reference Images | Visual context for subjects, products, layouts, or styles |
| Aspect Ratio | The final composition shape, including Auto and 10 fixed ratios |
Nano Banana is the efficient default. Choose Nano Banana Pro when the job needs professional asset production, more precise text, complex layouts, or higher-resolution delivery. Choose Nano Banana 2 when you want the newer high-efficiency generation and editing tier available in the Wizzx catalog.
Sources
Backend pricing
Rates published by the API.
Model identity, billing unit and every price tier below are read from the backend catalog.
1 credit = $0.001 USD. The final deduction returned by the API is authoritative.
Back to all models