Google launches Nano Banana 2 Lite at $0.034 per 1,000 images

Google's new Nano Banana 2 Lite generates images in 4 seconds at a fraction of the cost of its predecessors, targeting enterprise developers who need high-throughput visual generation.

Google's Nano Banana 2 Lite generates images in 4 seconds at $0.034 per 1,000 images, undercutting its predecessor by 13 percent while matching the quality of the more expensive Pro line. The model, technically designated Gemini 3.1 Flash-Lite Image, is available immediately through Google AI Studio, the Gemini API, and the Gemini Enterprise Agent Platform.

"This model is built for high-throughput commercial applications where speed and cost are the binding constraints," Google said in its technical documentation, positioning the Lite variant as an infrastructure workhorse rather than an artistic engine.

The model achieves a Text to Image arena Elo score of 1251 — surpassing the legacy Nano Banana 1 at 1151 and edging out the bulkier Nano Banana Pro at 1245. For editing tasks, it scores 1308 on single-image and 1294 on multiple-image benchmarks. The trade-off is resolution: Nano Banana 2 Lite supports only 1K output, while the standard Nano Banana 2 and Pro lines scale to 4K.

At $0.034 per 1,000 images, the Lite model costs roughly half the standard Nano Banana 2 at $0.067 and a quarter of Nano Banana Pro at $0.134. Google is betting that volume-driven enterprise workflows — programmatic ad generation, e-commerce product visuals, real-time A/B testing — will offset the lower per-unit revenue, while locking developers into its managed cloud stack.

A Price War in AI Image Generation

The pricing undercuts not only Google's own older models but also competitors in the rapidly commoditizing AI image generation market. Startup Krea recently released Krea 2 Turbo, which generates images in 2 seconds under a partially open license for small enterprises, though at an unspecified price point. OpenAI's ChatGPT image generation has driven significant user engagement, but the company has not disclosed per-image pricing for enterprise API access.

Nano Banana 2 Lite's capabilities include upgraded world knowledge for drafting data visualizations and contextual layouts, enhanced character consistency across sequential generations, and localized typographic rendering for English, Chinese, Japanese, and Korean text. These features target specific enterprise pain points: advertising teams running thousands of localized campaign variations, e-commerce platforms generating product scenes at scale, and design teams iterating on concept mockups.

Google also launched Gemini Omni Flash into public preview alongside Nano Banana 2 Lite. The multimodal model handles conversational video generation and editing, representing Google's longer-term bet on agentic video manipulation. While Omni Flash targets creative video workflows, Nano Banana 2 Lite is the immediate revenue driver — optimized for the high-volume, low-margin image generation that enterprises need today.

The release follows Google's decision on June 29 to make personalized AI image generation free for all eligible US users, removing the paywall that had restricted the feature to Plus, Pro, and Ultra subscribers since April. That feature, built on the original Nano Banana model, connects to users' Google account data across Gmail, Photos, and YouTube to generate context-aware images.

For investors, the question is whether Google's aggressive pricing can drive enough API volume to offset the per-unit margin compression. Alphabet trades at roughly 22 times forward earnings, with Google Cloud revenue growing at 30 percent year-over-year. If Nano Banana 2 Lite accelerates enterprise adoption of Google's AI platform, the cloud business — not image generation itself — stands to capture the most value.

This article is for informational purposes only and does not constitute investment advice.