Latest Google Vertex AI with enhanced capabilities
# Nano Banana Pro (G) Versatile AI model for high-quality creative and analytical tasks. ## Overview Nano Banana Pro (G) is a high-performance generative imaging model built on the advanced Gemini 3 Pro architecture. Designed specifically for the fast-paced requirements of modern marketing teams and advertising agencies, this model bridges the gap between raw creative ideation and production-ready visual assets. By leveraging deep contextual understanding, it excels at interpreting complex prompts to produce high-fidelity imagery that aligns with brand guidelines and campaign objectives. What sets Nano Banana Pro (G) apart is its integration with Google’s robust grounding technology, allowing the model to pull from real-world web context to ensure that generated visuals remain relevant to current trends and cultural moments. Whether you are creating static social media assets, high-resolution ad banners, or conceptual storyboards, this model provides the precision and flexibility required to maintain visual consistency across diverse marketing channels. ## Use Cases * Generating high-conversion social media ad creative for Instagram and LinkedIn. * Creating rapid visual prototypes for marketing campaign storyboards. * Producing high-resolution product photography mockups for e-commerce landing pages. * Developing consistent brand-aligned imagery for multi-channel digital advertising campaigns. * Enhancing existing creative assets through precise image-to-image editing and style transfer. * Visualizing abstract marketing concepts for internal stakeholder presentations. ## Parameters To achieve the desired output, users can configure several technical parameters including top_p, temperature, and aspect_ratio. Please refer to the input parameter configuration table in the Pixloop interface to adjust resolution, output format, and grounding settings before initiating your generation request. ## Tips for Best Results * Use descriptive, specific language in your prompt to guide the model toward your desired aesthetic; include lighting, camera angle, and color palette details. * Enable the 'grounding' feature when creating content related to current events or specific trending topics to ensure the model incorporates up-to-date visual context. * Experiment with the 'temperature' setting: lower values (0.5-0.8) are better for brand-consistent, predictable results, while higher values (1.2-1.5) are ideal for brainstorming unique, artistic concepts. * Utilize the 'input_images' parameter to provide style references, ensuring the output matches your brand's existing visual identity. * Keep the 'top_p' setting at the default 0.95 for most tasks to maintain a balance between creative diversity and logical coherence. * For high-impact digital displays, select the 4K 'image_size' option to ensure maximum clarity and detail in your final assets. ## About Nano Banana Pro (G) is powered by the Vertex AI infrastructure, utilizing the Gemini 3 Pro image-preview model lineage. Developed as a specialized iteration for the Pixloop platform, it benefits from Google's state-of-the-art multimodal research, combining massive-scale training data with refined fine-tuning to meet the professional demands of the advertising industry.
Model input reference derived from preset schema.
| Name | Type | Required | Default | Description |
|---|---|---|---|---|
| top_p | number | No | 0.95 | Cumulative probability threshold for token selection (0.0-1.0) |
| prompt | string | Yes | null | Text description of the image you want to generate or edit |
| grounding | boolean | No | false | Enable Google Search grounding for fresher web context. |
| image_size | select | No | "1K" | Output image resolution (Vertex imageConfig.imageSize). |
| temperature | number | No | 1 | Controls randomness in generation (0.0 = deterministic, 2.0 = maximum creativity) |
| aspect_ratio | select | No | "1:1" | Aspect ratio of generated images |
| input_images | image | No | [] | Optional reference images for editing or style transfer (up to 14 images) |
| output_format | select | No | "png" | Requested output format for generated images |
| candidate_count | select | No | "1" | Number of image variations to generate (fixed at 1 per request) |