Google releases Gemini 3.1 Flash Lite Image, its fastest and cheapest image generation model

Google has released Gemini 3.1 Flash Lite Image, also called Nano Banana 2 Lite, which the company describes as its fastest and cheapest image generation model. The model is available through Google's AI Studio and Gemini API with the identifier gemini-3.1-flash-lite-image.

According to Google, the model is "engineered for velocity and scale," positioning it as the entry-level option in their image generation lineup. The release follows Google's earlier Nano Banana models introduced in April 2026.

Model positioning

The "Lite" designation indicates this is Google's budget tier for image generation, optimized for speed and cost over quality. Google has not yet disclosed specific pricing per image or generation latency benchmarks.

The model is accessible through Google AI Studio's web interface and can be called programmatically via the Gemini API. This continues Google's strategy of offering multiple performance tiers across their Gemini model family.

Testing results

Early testing with complex prompts like "Where's Waldo"-style scenes shows the model can generate detailed, densely populated images with multiple characters and background elements. However, text rendering remains inconsistent—test images showed misspellings like "FOREE'S FESTIVAL" and "FOREST FIVAL" instead of "Forest Festival."

The model successfully handled compositional requirements, placing specific objects (like a raccoon with a ham radio) within crowded scenes, though text accuracy lags behind the visual composition capabilities.

What this means

Gemini 3.1 Flash Lite Image expands Google's image generation offerings to cover the low-cost, high-speed segment where developers need rapid iteration or high-volume generation at minimal cost. The "Lite" tier suggests Google is competing directly with other budget image models from Stability AI and others.

The persistent text rendering issues—a common weakness across image generation models—remain unresolved even in this latest release. Developers building applications requiring accurate text in images should continue to verify outputs or use alternative methods for text overlay.

Without disclosed pricing or benchmark comparisons, it's unclear how "Flash Lite" compares to competitors on cost-per-image or generation speed. Google's emphasis on "velocity and scale" suggests this targets high-throughput applications over quality-sensitive use cases.