Gemini 3.1 Flash Lite Image (Nano Banana Lite) reaches general availability

Google has moved gemini-3.1-flash-lite-image — nicknamed Nano Banana Lite — out of preview and into general availability, positioning it as its fastest, cheapest image generation and editing model in the Gemini API lineup.

What's new

The Gemini API changelog confirms Google "Released gemini-3.1-flash-lite-image (Nano Banana Lite) to general availability (GA), our built-in multimodal model optimized for ultra-low latency and cost-effective image generation and editing."

From the model card, the headline performance claim is sub-2-second end-to-end latency for a generation or edit request. Other specs:

• Resolution: 1024x1024px only — no 2K or 4K output (that headroom is reserved for the higher-tier Gemini 3.1 Flash Image and Gemini 3 Pro Image models)
• Aspect ratios: 14 supported, including 1:1, 3:2, 2:3, 3:4, 4:3, 4:5, 5:4, 9:16, 16:9, and 21:9
• Interleaved generation and editing: native support for both text-to-image and image+text-to-image workflows, aimed at fast multi-turn local edits like color swaps or background adjustments

Context

This GA release lands the same day as Gemini Omni Flash's video-generation preview, continuing Google's push to round out a full native-multimodal-generation stack inside the Gemini API. It follows the GA launch of gemini-3.1-flash-image (Nano Banana 2) and gemini-3-pro-image (Nano Banana Pro) in late May — those models handle higher-resolution, higher-fidelity image work, while Flash Lite Image is explicitly the low-latency, low-cost tier of that same family, built for interactive, high-volume, or cost-sensitive use cases rather than maximum output quality.

The positioning mirrors what Google has already done on the text side with its Flash and Flash-Lite model tiers: a full-capability version alongside a cheaper, faster-but-lower-ceiling sibling aimed at a different point on the cost/latency curve.

Why it matters

Sub-2-second image generation and editing at GA (not preview) status matters most for products that need image generation inline in a live user interaction — a chat app doing a quick avatar edit, a shopping app swapping a product color, a design tool iterating on a mockup — where a multi-second wait breaks the interaction. Capping resolution at 1024x1024 is a deliberate tradeoff: Google is betting that for a large share of interactive image-editing use cases, speed and cost matter more than maximum resolution, and that developers who need higher fidelity will reach for the Pro or full Flash Image tier instead. It's a concrete signal of how Google is segmenting its image-generation stack by latency and cost rather than shipping one model to do everything.