Gemini 3 Image Preview(即 Nano Banana)是一款先进的 AI 模型,专精于自然语言驱动的图像生成与编辑。它能生成超写实、符合物理规律且风格转换无缝的图像。

Model Type:
Pricing: Nano Banana API costs 4 credits per image (~$0.02). High-tier top-ups (+10% bonus) bring effective pricing down to ~$0.018 per image.
Input

The prompt for image generation

0/20000

Output format for the images

output ratio

A configurable parameter. Defaults to true in the Playground.

There is no guarantee that everything can be filtered out; if you are not satisfied with the results, you will need to make your own arrangements.
Output
output typeimage
Generated output
Examples

Explore different use cases and parameter configurations

README

Complete guide to using google/nano-banana

Nano Banana API & Nano Banana Pro API:高性价比 Gemini 图像模型集成

头像 1
头像 2
头像 3
头像 4
头像 5
5.0

from 5000+ reviews

Nano Banana Pro API 支持 64K 超大上下文窗口
Gemini 3 Pro Image API 提供 4K 高分辨率输出
使用 Nano Banana Pro API 实现多轮对话式编辑
基于 Gemini 3 Pro Image Preview 的 8 图组合能力
利用 Gemini 3 Pro Image API 实现基于搜索的精准生成
Nano Banana Pro API:实现完美的文字渲染
Gemini 3 Pro Image API:掌控影棚级画质
利用 Nano Banana Pro API 轻松驾驭复杂的多步骤工作流
借助 Gemini-3-Pro-Image-Preview 保持主体形象高度一致
利用 Nano Banana Pro API 本地化功能,实现设计全球化
直观的自然语言编辑体验

直观的自然语言编辑体验

Nano Banana API 只需简单的文本提示(比如“加个夕阳”或“把椅子换成王座”),即可实现高精度的图像编辑。它不仅能识别基础指令,更能精准解读复杂需求,将用户意图精准还原为视觉画面。

稳定可靠的输出质量

稳定可靠的输出质量

Gemini 2.5 Flash Image API 能够在多轮迭代编辑中保持高度连贯,有效避免画面失真或风格偏差。无论是对同一图像进行多次更新,还是批量应用相似的编辑操作,Nano Banana AI 图像编辑 API 都能确保输出结果的稳定性与一致性。

精准可控的图像编辑

精准可控的图像编辑

借助 Gemini 2.5 Flash Image API,物体替换、背景修改及面部优化等操作均能实现像素级的精准编辑。Nano Banana API 能够完好保留原始场景的细节,确保每一次修改都能无缝融合,呈现出专业级的图像品质。

极速性能

极速性能

速度是 Nano Banana API 的显著特征,Nano Banana AI 的图像生成与编辑仅需数十秒即可完成。这种极速性能完美支持从即时内容创作到快速原型设计的各类实时应用,让开发者在保证画质的同时,维持高效的工作流。

符合物理规律的逼真视觉效果

符合物理规律的逼真视觉效果

得益于 Gemini 先进的推理能力,Nano Banana API 能够生成符合现实物理规律的图像——呈现协调的光影效果、自然的阴影以及准确的空间关系。无论是处理复杂背景还是添加新物体,每一次编辑都能保持完美的视觉一致性。

5.0/ 5
50,688 ratings
Tap a star to rate
功能特性Nano Banana (Gemini 2.5 Flash Image)Nano Banana Pro (Gemini 3 Pro Image)Flux KontextQwen Image Edit (通义千问图像编辑)
Core FocusPrecise language-driven editing and scene consistencyStructured, multi-step image generation with 4K control & long-context promptsContextual image generation and layout coherenceFast, open-source image editing
Editing PrecisionStrong in object swaps, stable identityHigh precision in typography, layout, diagrams, and semantic editsModerate; better for broad style shiftsStrong; fast but less controlled
SpeedFast responses for quick tasksSlower than Nano Banana; optimized for multi-turn workflowsVariable; slower on complex tasksUltra-fast ~8s results
Realism & PhysicsHighly realistic textures, lighting, physicsStudio-quality realism, shot composition, controllable lightingContextual realism with inconsistenciesRealistic but may distort small details
Style & Design ControlVersatile across styles and rendering modesFine control over shot angle, lighting, typography, aspect ratioSmooth style shifts, artistic controlLimited by open-source framework
Multi-Image WorkflowsLimitedUp to 8 images blended with consistent lighting & perspectiveLimited multi-image supportLimited or inconsistent
Text & TypographyGood for short textClear, legible long text across languages; suitable for posters & UIStruggles with precise letteringCan generate text but lacks strict accuracy
ConsistencyMaintains identity across iterationsHigh-fidelity character and object consistency across scenesInconsistent on complex editsMay alter backgrounds/faces
Best Use CasesFast edits, lightweight creative tasksBranding packs, storyboards, packaging, infographicsStyle-driven creative conceptsQuick image edits, experiments
Pricing on Kie.ai~$0.02 per image~24 credits per image (≈ $0.12), ~20% cheaper than official price; subject to change$0.025 or $0.05 per image$0.0165 per image