AI Model
Nano Banana Pro Comfy Workflows
Run Google's Nano Banana Pro, the Gemini 3 Pro Image model, in ComfyUI: high-fidelity generation, multi-image fusion, and 4K output, on Comfy Cloud.
40 workflows
The workflows
40 Nano Banana Pro Comfy Workflows
About This AI Model
What is the Nano Banana Pro model?
Most people meet Nano Banana Pro comfyui workflows through one job: take several reference images and fuse them into one clean, high-fidelity result. Nano Banana Pro is Google DeepMind's flagship image model, the Gemini 3 Pro Image build, and it is a cloud model reached through ComfyUI's API nodes rather than a local checkpoint. The workflows on this page wire those calls into editable graphs, so the prompt, the reference images, and the output flow through nodes you can chain.
What makes it stand out is the combination of world knowledge and control. It can pull on Google's knowledge to render real-world subjects accurately, blend up to fourteen reference images into a single coherent output, and hold a character consistent across separate generations. Text rendering is a strength too, with clean lettering and translation across languages, and native output reaches 4K, up to 4096 by 4096, for work that has to hold up at size.
The family has a faster tier as well. Gemini 2.5 Flash Image, the original Nano Banana, trades some fidelity for speed and cost, which suits interactive editing loops, while Nano Banana Pro is the one to reach for when the result has to be production-ready. Both run through the same API nodes, so a workflow can target whichever fits the task.
Because these are open ComfyUI graphs around a hosted model, you control the inputs that shape the output: the prompt, the reference images, the aspect ratio, and how the result feeds the rest of your pipeline. To run one, open a Nano Banana Pro workflow on Comfy Cloud with nothing to install, add your prompt and reference images, and generate. Re-run with adjusted references until the fusion and detail land.
Reproducible, tunable, yours
Every step is exposed and adjustable, and the exact graph is yours to reuse across projects.
Start free. Upgrade when you’re ready.
Capabilities
What You Can Create
Nano Banana Pro is Google DeepMind's flagship image model, internally Gemini 3 Pro Image, accessed through ComfyUI API nodes. It fuses up to 14 references, renders 4K output, and holds character consistency.
World knowledge
World-knowledge grounding for accurate real-world subjects, plus clean text rendering.
Try workflow
Character consistency
Character consistency across separate generations, holding up to five people's identity.
Try workflow
WORKS
Connect models, processing steps, and outputs on a canvas where every decision is visible and every step is inspectable.
Start from a community template or build from scratch.
Comparison
Why Comfy Workflows
-
Fuse up to 14 references in one graph
Blend logos, product shots, color swatches, and style references into one coherent image, holding up to five people's identity across the result.
-
Pick your variant and settings
ONLY ONComfyCLOUDSwitch between Nano Banana Pro for fidelity and Gemini 2.5 Flash Image for speed, then set the prompt, reference images, and aspect ratio.
-
Native 4K that drops into your pipeline
Generate up to 4096 by 4096 and chain the result straight into other ComfyUI nodes for upscaling, compositing, or further editing.
-
Editable graphs around a hosted model
The Gemini 3 Pro Image calls run through ComfyUI API nodes you can wire, version, and re-run, with no watermark on the output.
Applications
What people use it for
How people use Nano Banana Pro for reference blends, edits, and 4K work.
Frequently Asked Questions
Nano Banana Pro FAQ
Keep exploring
Related models and use cases to try next.