Apple + Real ForgeFX Logo
Each square was generated by an image model from a prompt plus the official ForgeFX logo as an image reference. No output logo was pasted or composited afterward.
Exact base prompt
Create a square studio image of one crisp realistic red apple centered on a clean warm-white background. Put a large ForgeFX logo in the upper-right corner of the image, integrated naturally as a printed poster/sign/brand mark within the generated scene. Use the supplied ForgeFX logo reference image exactly: preserve the anvil icon, wordmark lettering, colors, proportions, and spacing as faithfully as the model can. Do not invent, redraw, stylize, retype, paraphrase, or approximate the logo. Do not add any other words. Do not leave the corner blank. The apple should be simple and secondary; the test objective is whether the model can render the supplied ForgeFX logo correctly inside the generated image.
Current documentation check
- OpenAI’s current GPT Image 2 model page says gpt-image-2 supports text and image input and image output.
- OpenAI’s image-generation guide says the Responses API supports flexible image inputs and image-generation/editing workflows.
- Google’s Nano Banana Interactions API docs, last updated 2026-06-30, say Nano Banana accepts text, images, and text+image combinations.
- Google’s Gemini 3 Pro Image docs say image input and output are supported, with up to 14 images per prompt.