Nano Banana 2 Lite API

Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image, model code gemini-3.1-flash-lite-image) is Google's lightweight model for fast image generation and editing. Positioned as the efficiency specialist of the image generation family, it targets sub-2-second end-to-end latency and significantly reduced TPU compute costs, serving high-volume interactive developer use cases and real-time consumer applications. The model launched in June 2026 (June 23, 2026 GA per Vertex AI), with native support for interleaved "Text → Text + Image(s)" and "Image + Text → Text + Image(s)" generation and editing, while maintaining the high character alignment of the original Nano Banana.

The model is offered through the iCreat platform as an API using a two-step asynchronous task flow: submit a task to obtain a task_id, then poll the result endpoint until a terminal status. Output resolution is fixed at 1K (1024×1024), billed per generated image at $0.04 per image; ten output aspect ratios from 1:1 to 21:9 are supported.

Model Positioning

Nano Banana 2 Lite targets image production where latency and cost come first. It is purpose-built for the 1K tier — without Nano Banana 2's 0.5K–4K resolution ladder or Nano Banana Pro's flagship fidelity — compressing per-image latency to under 2 seconds and per-image cost to the family's lowest, to power real-time experiences where users watch results arrive. Within Google's image generation family the division of labor is clear: Nano Banana Pro handles high-fidelity final assets, Nano Banana 2 handles everyday work needing references and size flexibility, and Nano Banana 2 Lite handles massive, latency-sensitive, cost-sensitive drafts and interactive editing. Typical needs include real-time previews, batch stickers and first drafts, and fast multi-turn local edits.

Core Capabilities

Sub-2-Second Latency (Official Target)​

Designed with a sub-2-second end-to-end latency target and substantially lower TPU compute costs than the family, it fits interactive applications where users wait on each generation and high-concurrency batch jobs.

Interleaved Generation and Editing

Natively supports "Text → Text + Image(s)" and "Image + Text → Text + Image(s)" — generate first, then keep editing based on the result in the same session, without switching endpoints.

Fast Multi-Turn Local Edits

Supports multi-turn local edits such as swapping colors, creating stickers, and adjusting backgrounds, while keeping character alignment on par with the original Nano Banana across edit rounds.

Reference-Image Guidance

The image array in the request receives reference image URLs for generation guidance or image editing; the output aspect ratio defaults to that of the first reference image.

Pricing

Billed per generated image only; the costUSD field in the response returns the actual cost once the task succeeds.

Application Scenarios

• Real-time interactive apps: in-chat image generation, live previews, and draft iteration
• Batch first drafts: high-volume, low-cost initial drafts for e-commerce listings and social media
• Fast multi-turn editing: recoloring, sticker making, and background tweaks in a lightweight pipeline
• Localized lightweight assets: operation graphics and placeholders where low resolution suffices
• Reference-driven re-creation: quickly deriving new visuals from reference images

Model Comparison

Same-Series Comparison

Cross-Model Comparison

Same-series data comes from Google's official documentation and release pages; this model's pricing follows the iCreat channel price table.

Why Choose Nano Banana 2 Lite?

• Sub-2-second end-to-end latency target — near-instant feedback for interactive scenarios
• Lowest per-image cost in Google's image generation family, keeping massive drafts and batch first drafts affordable
• Generation and editing interleaved in one flow, with no switching between separate endpoints
• Ten output aspect ratios (1:1 through 21:9) covering mainstream delivery sizes, defaulting to the first reference image's ratio
• Character alignment on par with the original Nano Banana, keeping text and subjects stable across multi-turn edits

Specifications

Architecture

Nano Banana 2 Lite belongs to the lightweight image branch of Google's Gemini 3.1 series (model code gemini-3.1-flash-lite-image), sharing its lineage with Nano Banana 2 (gemini-3.1-flash-image) but optimized for efficiency. The model natively supports interleaved multimodal input and output: text and images can form a conversation context in any order, and the model can return text and generated images in the same response, so generation and editing share one pipeline. It is specifically optimized for 1K (1024×1024) resolution; each input reference image consumes approximately 1,120 tokens per official documentation. Generated images carry a SynthID watermark (always on) plus C2PA content credentials.

Notes

• Submit and query must be chained with the same task_id
• While processing, result is []; keep polling, with a suggested interval of 2–5 seconds
• FAILED is terminal; check the prompt, reference image URLs, image format, and aspect ratio before resubmitting
• image_size supports 1K only; 2K/4K requests are rejected
• When aspect_ratio is not provided, it defaults to the first reference image's ratio
• Every item in the image array must be a publicly accessible image URL
• Tasks may fail when reference images are missing or inaccessible; ensure URLs are valid
• Output images carry a SynthID watermark (always on for Google; cannot be disabled)
• costUSD is returned only when the task succeeds; failed tasks do not produce a cost field

FAQ

How do I get started with the Nano Banana 2 Lite API?​

Register on the iCreat platform and obtain an API Key from the console, send a generation request to the submit endpoint, then poll the result endpoint with the returned task_id. All requests are authenticated with the Authorization: Bearer header. Keep your API Key safe and never expose it in client code or public repositories.

How is the cost calculated?​

Billed per generated image: $0.04 per image at 1K resolution. For example, generating 10 images costs $0.40. The costUSD field in the response gives the actual cost once the task succeeds.

Does it support generation or editing?​

Both. The same submit endpoint handles text-to-image and image editing: with only a prompt, it generates from the prompt; with the image reference array, it generates from references or applies prompt-described edits. For multi-turn editing, submit the previous result as the next reference image.

What resolutions and aspect ratios are supported?​

Resolution supports 1K (1024×1024) only — 2K and 4K are not supported. Aspect ratios support ten values: 1:1, 2:3, 3:2, 3:4, 4:3, 4:5, 5:4, 9:16, 16:9, and 21:9; when unspecified, the default follows the first reference image's ratio. Pick the right size for each delivery platform.

How do I query the task result?​

Send a POST request to https://api.icreat.ai/v1/task/result with the task_id in the body. A status of SUBMITTED means processing, with result as [] — keep polling. When status is SUCCEEDED, result returns the image resource array; read url or download_url to retrieve the image.

What if the task fails?​

FAILED is a terminal status. Check in order: whether the prompt is empty, whether reference image URLs are publicly accessible, whether image_size is 1K, and whether aspect_ratio is one of the ten allowed values. Fix any issue and resubmit. The cost field is returned only when the task succeeds.