Skip to main content
Endpoint: POST https://fal.run/fal-ai/nano-banana-2 Endpoint ID: fal-ai/nano-banana-2

Try it in the Playground

Run this model interactively with your own prompts.

​
Quick Start

import fal_client

def on_queue_update(update):
    if isinstance(update, fal_client.InProgress):
        for log in update.logs:
           print(log["message"])

result = fal_client.subscribe(
    "fal-ai/nano-banana-2",
    arguments={
        "prompt": "An action shot of a black lab swimming in an inground suburban swimming pool. The camera is placed meticulously on the water line, dividing the image in half, revealing both the dogs head above water holding a tennis ball in it's mouth, and it's paws paddling underwater."
    },
    with_logs=True,
    on_queue_update=on_queue_update,
)
print(result)

​
Input Schema

​
prompt
string
required
The text prompt to generate an image from.
​
num_images
integer
default:"1"
The number of images to generate. Default value: 1Range: 1 to 4
​
seed
integer
The seed for the random number generator.
​
aspect_ratio
Enum
default:"auto"
The aspect ratio of the generated image. Supports extreme ratios: 4:1, 1:4, 8:1, 1:8. Use “auto” to let the model decide based on the prompt. Default value: autoPossible values: auto, 21:9, 16:9, 3:2, 4:3, 5:4, 1:1, 4:5, 3:4, 2:3, 9:16, 4:1, 1:4, 8:1, 1:8
​
output_format
OutputFormatEnum
default:"png"
The format of the generated image. Default value: "png"Possible values: jpeg, png, webp
​
safety_tolerance
SafetyToleranceEnum
default:"4"
The safety tolerance level for content moderation. 1 is the most strict (blocks most content), 6 is the least strict. Default value: "4"Possible values: 1, 2, 3, 4, 5, 6
​
sync_mode
boolean
default:"false"
If True, the media will be returned as a data URI and the output data won’t be available in the request history.
​
resolution
ResolutionEnum
default:"1K"
The resolution of the image to generate. Default value: "1K"Possible values: 0.5K, 1K, 2K, 4K
​
limit_generations
boolean
default:"true"
Experimental parameter to limit the number of generations from each round of prompting to 1. Set to True to to disregard any instructions in the prompt regarding the number of images to generate and ignore any intermediate images generated by the model. This may affect generation quality. Default value: true
​
enable_web_search
boolean
default:"false"
Enable web search for the image generation task. This will allow the model to use the latest information from the web to generate the image.
​
thinking_level
Enum
When set, enables model thinking with the given level (‘minimal’ or ‘high’) and includes thoughts in the generation. Omit to disable.Possible values: minimal, high

​
Output Schema

​
images
list<ImageFile>
required
The generated images.
​
description
string
required
The description of the generated images.

​
Input Example

{
  "prompt": "An action shot of a black lab swimming in an inground suburban swimming pool. The camera is placed meticulously on the water line, dividing the image in half, revealing both the dogs head above water holding a tennis ball in it's mouth, and it's paws paddling underwater.",
  "num_images": 1,
  "aspect_ratio": "auto",
  "output_format": "png",
  "safety_tolerance": "4",
  "sync_mode": false,
  "resolution": "1K",
  "limit_generations": true,
  "enable_web_search": false
}

​
Output Example

{
  "images": [
    {
      "content_type": "image/png",
      "file_name": "nano-banana-2-t2i-output.png",
      "url": "https://storage.googleapis.com/falserverless/example_outputs/nano-banana-2-t2i-output.png"
    }
  ],
  "description": ""
}

​
Nano Banana 2 [text-to-image]

Google’s Gemini 3.1 Flash Image architecture generates vibrant, high-fidelity visuals at speed, combining the reasoning capabilities of a multimodal foundation model with the efficiency of Flash-optimized inference. It understands creative intent holistically rather than matching keywords, producing images with accurate text rendering, character consistency, and coherent spatial composition in seconds. Built for: Marketing campaigns and social media assets | Product photography and visualization | Designs requiring accurate in-image typography | Storyboarding with consistent characters across frames

​
Reasoning-Guided, Flash-Fast

Built on Google’s Gemini 3.1 Flash Image foundation, Nano Banana 2 reasons about composition, lighting, and spatial relationships before rendering. Unlike traditional diffusion models that treat prompts as weighted tokens, this architecture interprets creative direction as a multimodal language model would, capturing nuance and context that single-modality systems miss, then executes at Flash-tier speed. What this means for you:
  • Vibrant output: Rich color, punchy contrast, and visual coherence out of the box without post-processing
  • Accurate text rendering: Character-by-character validated typography in multiple languages, directly in generated images
  • Character consistency: Maintain identity for up to 5 people across generations for storyboarding and campaign work
  • Natural language control: Describe mood, style, and context conversationally without mastering prompt engineering syntax
  • Web-grounded generation: Optionally ground outputs in real-time web information for factually current visuals

​
Technical Specifications

SpecDetails
ArchitectureGemini 3.1 Flash Image (Nano Banana 2)
InputText prompts (natural language); up to 14 reference images for editing
Output FormatsPNG, JPEG, WebP
Resolution1K (default), 2K (1.5x rate), 4K (2x rate), 512x512 (0.75x rate)
Aspect Ratiosauto, 21:9, 16:9, 3:2, 4:3, 5:4, 1:1, 4:5, 3:4, 2:3, 9:16
Batch1-4 images per request
Character ConsistencyUp to 5 people across generations
WatermarkingSynthID digital watermarking on all outputs
Web SearchOptional grounding via enable_web_search or enable_google_search
LicenseCommercial use enabled through fal.ai
API Documentation

​
How It Stacks Up

vs. Nano Banana Pro (Gemini 3 Pro Image): Nano Banana 2 prioritizes speed and vibrant output on the Flash architecture, generating in seconds where Pro optimizes for maximum reasoning depth at $0.15/image. Choose Nano Banana 2 for fast iteration and production volume, Pro for maximum compositional complexity. vs. FLUX.2 [dev]: Nano Banana 2 delivers semantic-aware generation with native text rendering and character consistency through Gemini’s multimodal reasoning. FLUX.2 [dev] prioritizes resolution control and fine detail preservation for technical illustration workflows. vs. Original Nano Banana (Gemini 2.5 Flash Image): Nano Banana 2 adds reasoning-guided generation, dramatically improved text rendering, native multi-resolution output (1K/2K/4K), character consistency, multi-image compositing, and web search grounding. A generational leap in quality while maintaining Flash-tier speed.

​
Limitations

  • num_images range: 1 to 4
  • output_format restricted to: jpeg, png, webp
  • safety_tolerance restricted to: 1, 2, 3, 4, 5, 6
  • resolution restricted to: 0.5K, 1K, 2K, 4K
  • thinking_level restricted to: minimal, high