Google on February 26, 2026 announced Nano Banana 2, the latest version of its widely used image generation model. The company says the upgrade — technically branded as Gemini 3.1 Flash Image — produces more realistic results than the prior release.
Nano Banana 2 is also becoming the default image model inside the Gemini app, spanning its Fast, Thinking, and Pro modes. The change signals Google’s intent to make “good enough, fast” image generation the baseline experience for most users.
Google first introduced Nano Banana in August 2025, and the model quickly took off inside the Gemini app — with especially heavy usage reported in markets like India. In November 2025, Google followed up with Nano Banana Pro, positioning it as the higher-quality option for more detailed generations.
What’s new in Nano Banana 2
Google says Nano Banana 2 keeps some of the high-fidelity traits of the Pro model while delivering results faster. The company claims it supports outputs from 512px up to 4K, along with multiple aspect ratios.
More consistency, more complexity
One of the headline improvements is better continuity for storytelling-style prompts. According to Google, Nano Banana 2 can maintain character consistency for up to five characters and keep fidelity across up to 14 objects in a single workflow — useful for multi-scene concepts or iterative creative work.
Google also says the model handles more nuanced prompts, and can deliver more vibrant lighting, richer textures, and sharper detail compared with earlier versions.
Where it’s rolling out
With the launch, Nano Banana 2 is becoming the default image generator across the Gemini experience. Google also says it will become the default image model inside its video editing tool Flow — language worth watching as the change propagates, since defaults can vary during staged rollouts.
In Search, Google says Nano Banana 2 will become the default for image-related results and experiences, including Google Lens and AI Mode, rolling out across 141 countries on the Google app and on the web (desktop and mobile).
What Pro and paid users get
Google’s higher-tier plans — Google AI Pro and Ultra — will still allow subscribers to use Nano Banana Pro for specialized tasks. The company says users can access Pro by regenerating images via the three-dot menu, keeping the higher-fidelity option available when it matters.
Developer access and watermarking
For developers, Google says Nano Banana 2 will be available in preview via the Gemini API, Gemini CLI, and Vertex AI, and also through AI Studio and its development tool Antigravity (released last November).
Google also says every image generated by the new model will include a SynthID watermark to indicate AI-generated media. The company adds that the output supports C2PA Content Credentials interoperability, and claims SynthID verification inside the Gemini app has been used more than 20 million times since it launched there in November.