Nano Banana 2
Next-generation image synthesis on WaveSpeed — sub-second inference, built-in typography, and prompt-driven editing in a single model.
One Model, Three Capabilities
Nano Banana 2 combines generation, typography, and editing into a single architecture — optimized end-to-end on WaveSpeed infrastructure for maximum throughput.
Sub-Second Generation
Powered by WaveSpeed's optimized inference stack, Nano Banana 2 delivers high-fidelity images in under one second. Built for real-time workflows, interactive products, and production-scale pipelines.
Pixel-Perfect Typography
Reliable, legible text baked directly into generated images. Whether it's headlines on posters, labels on packaging, or UI copy in mockups — the output reads exactly as prompted.
Prompt-Driven Editing
Edit any image with plain language instructions — no masking, no inpainting tools. Describe what you want changed and the model handles composition, lighting, and context automatically.
Examples
Red-haired girl with porcelain skin and freckles posing with a blue hyacinth macaw on her shoulder, deep blue studio backdrop, fine art photography, shallow depth of field.
Two silhouettes standing in an enchanted forest gazing up at giant bioluminescent planets glowing with coral reefs and ocean life, painted sky at twilight.
A lone figure facing a colossal crystalline dragon made of molten gold and amber light, scorching orange desert sky, epic fantasy concept art, volumetric lighting.
Elderly man in a white linen suit carrying an enormous iridescent fish on his back with thin ropes, soft overcast light, muted pastel tones, magical realism.
Start Building
Integrate Nano Banana 2 with a single API call. Python, JavaScript, or cURL — ship in minutes.
- Optimized inference — no cold starts, instant scale-up
- Usage-based pricing — pay only for what you generate
- Python & JavaScript SDKs + REST API
FAQ
Nano Banana 2 is WaveSpeed's latest image generation model, purpose-built for speed without compromising quality. It handles text-to-image generation and prompt-driven editing through a single unified architecture.
Version 2 brings significantly faster inference, built-in typography rendering, prompt-based editing without masks, and native high-resolution output — all running on WaveSpeed's latest inference infrastructure.
The model supports a range of resolutions and aspect ratios including 1:1, 16:9, 9:16, 4:3, and 3:4 — from standard web sizes up to high-resolution print-ready output.
You provide an input image and a text instruction describing the change. The model interprets spatial context, handles blending, and applies the edit — no manual masks or selection tools needed.
Nano Banana 2 uses WaveSpeed's standard pay-per-use pricing. You only pay for what you generate — visit the pricing page for current rates and volume tiers.
Yes. All images generated through Nano Banana 2 are cleared for commercial use under WaveSpeed's standard terms of service.