Details
- Google AI Developers announces that Nano Banana Pro (gemini-3-pro-image) and Nano Banana 2 (gemini-3.1-flash-image) are now generally available for production use through the Gemini API.
- Both models are positioned as image-capable variants within the Gemini 3 family, aimed at developers needing vision and multimodal features in production workflows.
- Google highlights community-built demos showing advanced world knowledge, indicating these models can interpret and reason about complex visual scenes and related context.
- Additional examples emphasize subject consistency, suggesting the models can maintain the same character or object across multiple generated images.
- Style transfer examples showcase the ability to apply different visual aesthetics to content, while text rendering demos underline improved handling of written text inside images.
- A shared "production-ready specs" resource outlines performance, latency, and recommended use cases, signaling that the models have cleared internal readiness thresholds for enterprise deployment.
- Google directs developers to start building immediately via a "Start building today" link, funneling users to Gemini API documentation and quickstart resources for integrating the new models.
Impact
By pushing Nano Banana Pro and Nano Banana 2 to general availability, Google strengthens its Gemini vision stack just as rivals like OpenAI and Anthropic expand their own multimodal offerings. Exposing production-ready image models with clear specs and community examples lowers the barrier for developers to adopt advanced visual reasoning, potentially accelerating multimodal app development across creative, design, and enterprise scenarios.
