Google reveals Nano Banana 2 AI image model, coming to Gemini today
The new Gemini 3.1 Flash Image model can render 14 objects consistently and outputs up to 4K resolution.
Google has announced Nano Banana 2, the latest iteration of its popular AI image generation model, now officially branded as Gemini 3.1 Flash Image. This release consolidates Google's image generation offerings, replacing both the standard and Pro variants of the original Nano Banana across the Gemini app, Search, AI Studio, Vertex AI, and Flow. The company positions this as a significant upgrade, promising the advanced visual fidelity and world knowledge of the Pro model but delivered at the faster inference speeds previously associated with the Flash tier. This move signals strong confidence in the new model's capabilities, making it the sole option for image generation within Google's ecosystem.
Technically, Nano Banana 2 is built on the Gemini 3.1 large language model, which provides enhanced world knowledge pulled from the internet, leading to greater rendering fidelity and more accurate infographics. Google highlights specific improvements: the model can maintain character consistency for up to five figures at a time and accurately render as many as 14 distinct objects within a single workflow. It also boasts professional-grade text accuracy within images, richer textures, and more vibrant lighting. Output flexibility is increased with a wider range of aspect ratios and resolutions, scaling from a 512px square up to 4K widescreen formats. For users, this means a single, faster model capable of handling complex visual storytelling and detailed infographic creation, potentially streamlining workflows for developers and creatives using Google's AI tools.
- Replaces both Nano Banana and Nano Banana Pro as the default image model across all Google AI platforms.
- Delivers Pro-model quality with Flash-model speed, powered by the Gemini 3.1 LLM for better world knowledge.
- Can consistently render up to 14 objects per scene and output images in resolutions up to 4K widescreen.
Why It Matters
Consolidates Google's image AI into a single, faster, more capable model for developers and creatives, simplifying tool choice and improving output quality.