Google launches Nano Banana 2 Lite! Generates images in 4 seconds, only $0.034 per thousand images, targeting the enterprise automated AI image market.

Generative AI Image Battle Enters the "Cost and Speed" Close Combat! Google officially launched a streamlined image generation model for enterprises today (30th), called Nano Banana 2 Lite (also known as Gemini 3.1 Flash-Lite Image). It emphasizes ultra-low cost and high performance, costing only $0.034 per 1k images generated, and can produce a standard 1K image within 4 seconds, strongly targeting the programmatic advertising and automated workflow market.
(Previous: Google Launches the Latest AI Image Generation Tool Nano Banana Pro! Official Prompt Demonstrates Four Highlights)
(Background: Google's Strongest Image Generation Model Gemini 2.5 Flash Image (Nano Banana) Free to Use, Amazing Effects at a Glance)

Table of Contents

Toggle

• Sacrificing High Quality for Extreme Speed, Precisely Targeting Three Major Enterprise Pain Points
• Google Creative Model Family Pricing Comparison: Breaking the Cost Floor
• Simultaneous Unveiling of Multimodal Video Model Omni Flash

As enterprise demand for generative AI shifts from "experimental demonstrations" to "high-frequency practical applications," computing costs and generation latency have become the biggest barriers to adoption. To address this pain point, Google announced on June 30, 2026, the launch of the new Nano Banana 2 Lite model (also referred to as Gemini 3.1 Flash-Lite Image).

Google emphasized that this model is not designed for artists pursuing extreme aesthetics, but rather as a "high-throughput infrastructure workhorse" tailored for software engineers, programmatic advertising platforms, and digital commerce applications.

Sacrificing High Quality for Extreme Speed, Precisely Targeting Three Major Enterprise Pain Points

Nano Banana 2 Lite is built on the Gemini 3.1 Flash Lite architecture. To compress generation speed to within 4 seconds, Google strategically capped its maximum resolution at 1K canvas (not supporting the 2K/4K quality of standard NB2 or NB Pro). However, it has seen significant upgrades in several core capabilities that enterprises value most:

• Enhanced Character Consistency: Maintains high consistency of objects and character appearances in continuous image streams, storyboards, or e-commerce digital try-ons.
• Improved Text Rendering: Can generate clear, readable text within images, especially suitable for rapid generation and real-time validation of multilingual ad layouts.
• Rich World Knowledge: Can quickly generate accurate data visualization charts and location-specific context mockups.

In performance benchmarks, this streamlined model performed impressively. Its Text-to-Image Elo score reached 1251 points, not only surpassing the previous generation NB1's 1151 points but also slightly exceeding the high-end version NB Pro's 1245 points; its scores for single and multiple image editing also reached 1308 and 1294 points respectively.

Google Creative Model Family Pricing Comparison: Breaking the Cost Floor

This new model adopts an enterprise-first closed API model. While it cannot be deployed locally like some open-source models, it eliminates the heavy burden of hardware management for enterprises. Its most disruptive weapon is undoubtedly the extremely low pricing strategy:

| Model Version | | --- | Pricing (per 1k images) | Core Positioning & Features | | --- | --- | --- | | Nano Banana 2 Lite | $0.034 | Latest streamlined version. Fastest speed, lowest cost, suitable for large-scale automated production. | | Old Version NB1 | $0.039 | Previous generation base model, performance and speed gradually being replaced. | | Standard Version NB2 | $0.067 | Balanced option, supports higher resolution and details. | | High-End Version NB Pro | $0.134 | Flagship quality, highest cost, suitable for high-quality art and creative visual creation. |

Google's internal assessment indicates that Nano Banana 2 Lite provides about 60–70% of the general capabilities of high-end models, but at a significantly reduced cost. By tightly integrating with Google Workspace and the AI product line, Google hopes to use this low-price strategy to lock enterprise workflows into the Google ecosystem, countering competitors like Krea 2 Turbo.

Simultaneous Unveiling of Multimodal Video Model Omni Flash

Currently, Nano Banana 2 Lite is available for immediate use by developers through Google AI Studio, the Gemini API, and the Gemini Enterprise Agent Platform (GEAP). Notably, Google also released a public preview of Gemini Omni Flash today, a multimodal conversational video generation and editing model, demonstrating Google's powerful AI layout covering both static images and dynamic video.