Skip to main content

​
Overview

The Nano Banana series includes five image generation models, all available through APIYI at prices significantly lower than official rates. This page summarizes all Nano Banana model pricing to help you choose the best billing plan.
When calling the API, the token type you use determines the billing method. Nano Banana Pro, 2.1, 2, and Lite all support two billing modes: per-request and token-based. Each has two tiers: default stable tier and enterprise high-availability tier.

​
Model Overview

​
Nano Banana Pro Pricing

Model: gemini-3-pro-image-preview

​
Per-request Billing

$0.09 per request, unified pricing for 1-4K resolution. Simple billing, great savings on 4K images.
Save up to 68.75% on 4K images! Per-request billing doesn’t differentiate by resolution — 4K costs the same as 1K, so higher resolution means better value.

​
Token-based Billing (Not yet open, beta available)

Estimated per-image cost:
Token-based costs vary with input/output image tokens. These are estimates. Token-based billing is better for 1-2K images; per-request billing is more cost-effective for 4K.

​
Nano Banana 2.1 Pricing

Model: gemini-nano-banana-2.1 (released October 6, 2026; an upgrade to Nano Banana 2 — see Nano Banana 2.1 Image Gen/Editing)

​
Per-request Billing

$0.05 per request, the same price for 1K–4K, so 4K images get the best value.

​
Token-based Billing

Measured cost per image with token-based billing (2026-10-07, including default thinking tokens):
Nano Banana 2.1 thinks by default, so each image adds roughly 400–1300 thinking tokens billed at the output rate together with the image, so the token-based cost per image varies. Token-based is cheaper for 1K / 2K; per-request ($0.05) is cheaper for 4K. 512px is not supported.
Model prices are aligned with the official website and may change with it; the table above is for reference only — the Model Pricing tab in the top navigation is authoritative: Model Pricing.

​
Nano Banana 2 Pricing

Model: gemini-3.1-flash-image-preview

​
Per-request Billing

$0.055 per request, unified pricing for 1-4K resolution.

​
Token-based Billing

Estimated per-image cost:
These are estimates based on actual input/output tokens. Stack with top-up bonuses (up to 20%) for even lower costs.

​
Nano Banana 2 Lite Pricing

Model: gemini-3.1-flash-lite-image The fastest, cheapest tier — focused on the 1K canvas, ~4s per image. Prefer Pay-as-you-go Priority: the real-world unit price is lower and predictable. Estimated per-image cost:
Predictable in practice: a 1K image under token-based (per-token) billing averages ~$0.018/call in practice (typical range ~$0.016–$0.019, driven by output tokens), cheaper than the flat per-call $0.025/image. Stack with top-up bonuses for even lower costs.

​
Per-request Billing

Which billing mode? For Lite, prefer Pay-as-you-go Priority — ~$0.018/call in practice, cheaper than per-call $0.025 and predictable, and one token also covers Nano Banana Pro / 2’s per-call billing to run the whole series. The per-call $0.025/image is a flat rate that may be lowered later but is unchanged for now.

​
Nano Banana Gen 1 Pricing

Model: gemini-2.5-flash-image 1K images only, fixed per-request billing. Also eligible for top-up bonus promotions.

​
Top-up Bonuses

Stack top-up bonuses on top of already discounted prices for even greater savings.

Top-up Bonus Details

View current top-up bonus tiers and promotion details.

Token Management

Create and manage your API tokens.

​
Token Type Selection Guide

When creating a token, the “Billing model” setting determines the billing method. Choose based on your primary use case:
Simple rule: If you mainly generate 4K images, choose per-request billing (same price regardless of resolution). For mixed resolution usage, token-based billing saves more on lower resolutions.

​
A Changing Bill Is Not a Price Change

Two bills for the same model can differ a lot without any price change: under token-based billing the cost of a single request floats with your parameters. The two biggest measured drivers: A request costing about $0.033 yesterday and about $0.052 today on the same model usually means different parameters, not a price adjustment.
Per-tier measurements (tokens, cost, latency) are in Nano Banana 2 Image Gen/Editing and Usage Fields & Output Explained.
Choose per-request billing when you need predictable cost: one price from 1K to 4K (Pro $0.09 per image, Nano Banana 2 $0.055), independent of resolution and token counts.