Hedra
  • Pricing
Docs
Get started
Hedra
Get started
A flower-shop portrait with natural light and abundant detail.

Nano Banana 2

Multi-subject stills up to 4K — hand it several references and it keeps each one recognizable.

Get an API keyTry in Playground
All models

Overview

Nano Banana 2 (Gemini 3.1 Flash Image) is a high-speed image generation and editing model developed by Google. Built for rapid iteration, it delivers the advanced reasoning and real-world knowledge of Nano Banana Pro at significantly faster speeds. The model is especially well-suited for producing 4K-resolution assets, rendering accurate text for infographics and marketing mockups, and maintaining precise subject consistency across complex edits.

Specifications · Text + image to image

Input mode
Text + image to image
Accepts
reference image (up to 14)
Aspect ratios
adaptive, 16:9, 9:16, 1:1, 21:9, 4:3, 3:4, 3:2, 2:3, 5:4, 4:5, 1:4, 4:1, 1:8, 8:1
Resolutions
512px, 1K, 2K, 4K

Pricing

Google
512px6¢/generation
1K8¢/generation
2K12¢/generation
4K16¢/generation
plus $0.50 per 1M input tokens (each input image counts as 1120 tokens)

Build with this model: the Nano Banana 2 API on the Hedra Developer Platform.

Nano Banana 2 API →

Real output · generated with Nano Banana 2

Best of Nano Banana 2

A flower-shop portrait with natural light and abundant detail.Flower-Shop Window Portrait — Nano Banana 2Photoreal mid-century modern walnut accent chair with tan leather upholstery in a sunlit Scandinavian living room with linen curtains, oak flooring and a fiddle-leaf figWalnut Accent Chair, Scandinavian Interior — Nano Banana 2A close-up, square 2048x2048 image generated by Nano Banana 2 showing a fluffy orange and white tabby cat eating seared, bite-sized steak pieces from a small ceramic plate. The scene is set on a rustic wooden table indoors, with soft natural light coming from a window in the background.Ginger Cat Eating Steak, by Nano Banana 2A grid of ten numbered images showing a grey-haired man from behind. In each panel, he wears a different colored tactical backpack and jacket, set in various environments like a snowy cabin, a study, or a desert. Generated using Nano Banana 2 at 5504x3072 resolution.Ten Adventurer Backpack Variations, by Nano Banana 2A high-resolution 5504x3072 image generated by the Nano Banana 2 model. It shows a clear glass perfume bottle labeled AETHEREA Romano filled with amber liquid, topped with a gold cap. The bottle sits on a white marble table in the warm sunlight, with a blurred historic Italian street in the background.Luxury Perfume Bottle in Rome, by Nano Banana 2

Prompting

Prompt tips

  • Supply raw data for infographics: Paste your raw notes or statistics directly into the prompt and ask the model to convert them into a structured flat-lay infographic or diagram.
  • Force aspect ratios with follow-ups: If the model ignores your initial aspect ratio request, use a follow-up prompt in the same session to remind it of a specific supported ratio (e.g., '16:9').
  • Request in-image translations: Leverage its multilingual text capabilities by explicitly asking the model to localize signage or text elements (e.g., 'translate the storefront sign to Japanese').
  • Generate video base layers: Create your initial text-accurate, widescreen (16:9) assets here, then export them to animate with Veo 3.1.

About the model

Questions, answered

What is Nano Banana 2 best used for?

Nano Banana 2 is built for generating legible text, infographics, and diagrams. Because it connects to Google's real-world knowledge base, it can accurately render specific subjects and turn notes into structured visuals. The community frequently uses it for multi-image composition, such as style transfers, storyboarding, and maintaining character consistency across different scenes. It also supports video files as input prompts to analyze visual context and generate related images like thumbnails.

What is the lineage of Nano Banana 2?

Google announced Nano Banana 2 (officially known as Gemini 3.1 Flash Image) on February 26, 2026. It is the direct successor to the original Nano Banana, which was based on the Gemini 2.5 Flash architecture. This update brings the reasoning and text-rendering capabilities of the larger Nano Banana Pro into a faster model designed for high-volume, low-latency tasks.

How can I get the best results with Nano Banana 2?

Take advantage of the model's multimodality by uploading multiple reference images for style transfers or character consistency. You can also use video files as input prompts, allowing the model to analyze the footage and generate context-aware images. For text prompts, Google recommends specifying the characters, setting, and overall mood. You can find more techniques in Google's official prompting guide.

Same API · one key

Similar models

The felt robot paddling through a handmade lily pond.Nano BananaGoogleAn exploded burger with five deliberately ordered layers.Nano Banana ProGoogleCut-paper seahorse with layered coral and tactile shadows.Dreamina 3.1ByteDanceChrome and leather headphones with controlled studio reflections.Flux 1.1 ProBlack Forest LabsA fox mid-leap: motion, fur and fine snow detail.Flux 1.1 UltraBlack Forest LabsA miniature expedition climbs a towering cauliflower landscape.Seedream 5.0 LiteByteDance

Build with Hedra

Get an API keyTry in Playground

Research expertise and GPU infrastructure for models you own.

© 2026 Hedra

Products

  • Inference
  • Model training
  • GPU Sandboxes
  • World Model Runtime
  • Pricing

Deployment

  • Deployment options
  • Sovereign deployment

Model Factory

  • Overview
  • Forward Deployed Researchers
  • Enterprise

Research

  • Overview
  • Recursive self-improvement
  • Mechanistic interpretability
  • Post-training
  • Inference
  • Blog

Applications

  • Physical AI
  • Defense
  • Media generation

Developers

  • Platform onboarding
  • SDKs & integrations
  • Documentation
  • Model APIs
  • MCP & CLI
  • Changelog

Company

  • About Hedra
  • Manifesto
  • Careers
  • Contact
  • Play
  • Play pricing
  • Play help
  • Status
  • LinkedIn
  • X
  • GitHub
  • Discord
  • Privacy Policy
  • Terms of Use
  • Acceptable Use
  • Cookie Policy
  • Biometric Data Policy
  • Security