Nano Banana 2.1 lands on AI Gateway: product recontextualization, mask-based edits and factuality all improve
Synopsis
Google's Nano Banana 2.1 is now available on AI Gateway as google/gemini-nano-banana-2.1, so developers can call it directly. Compared with earlier Nano Banana models, it improves image generation and editing, covering product recontextualization, mask- and ink-based editing, and factuality, while maintaining Flash-tier latency and cost. It can be used through the AI SDK, the OpenAI-compatible Chat Completions API, or AI CLI.
Interpretation
Google's Nano Banana 2.1 is offered on AI Gateway as google/gemini-nano-banana-2.1, letting developers call this image generation and editing model without standing up their own service. Compared with earlier Nano Banana models, this version improves image generation and editing, specifically product recontextualization, mask- and ink-based editing, and factuality. The improvements are given as capability categories: placing products in new scenes, limiting edits to masked or marked regions, and depicting real-world subjects more accurately.
The model renders photorealistic skin tones, detailed materials, sharp lighting, and coherent backgrounds while maintaining Flash-tier latency and cost. The entry point is AI Gateway; after setting AI_GATEWAY_API_KEY, the model is used through the AI SDK, the OpenAI-compatible Chat Completions API, or AI CLI. The usage example passes the model name and a prompt through the AI SDK's generateText to produce a studio photo of a ceramic coffee mug on a white background.
Perspective
The report covers the model's availability on AI Gateway, the stated capability categories, and the call paths; improvements are given as capability descriptions without benchmark comparisons or quantitative metrics.
Improvements are stated as capability categories without specific comparison data against earlier Nano Banana models; Flash-tier latency and cost are not given as concrete figures.