Image generation powered by OpenAI's gpt-image-2.5-flare (GPT Image 2.5 Flare), the fastest model in the GPT Image 2.5 family, positioned for high-quality everyday image generation: it renders noticeably faster than Sunburst while keeping next-generation detail, and handles photo-realistic scenes, illustrations, product shots and many design styles. Describe the picture you want in a few sentences and receive a high-resolution output quickly; style, aspect ratio and subject are all controllable via prompt. A quality selector above the input box defaults to auto (the model decides from the prompt) and also offers low, medium, high, xhigh and max; higher levels render more detail and take longer. You can also upload up to sixteen pictures and edit them with plain-language instructions such as "replace the background with a beach" or "make the cup red"; if the result is not quite right, just send another instruction and the previous output is reused for continuous refinement, while uploading a new picture starts from that picture instead. Use Flare for fast iteration and bulk drafts, then switch to Sunburst for the final precision edit. The generated image size is chosen automatically by the model from your prompt: commonly 1024x1024 (square), 1024x1536 (portrait), or 1536x1024 (landscape), and other resolutions are possible, with the longest edge up to 3840 pixels. Maximum total upload size per request: 256MB (all files and message content combined). Accepted image formats: jpeg, jpg, png, webp. Maximum number of images: 16. Maximum size per image: 50MB. Image dimensions: at least 200 pixels per side. Maximum image resolution: 36,000,000 pixels in total. Output content formats: Image Image output size: chosen automatically by the model from the prompt; commonly 1024x1024, 1024x1536, or 1536x1024 pixels, and other resolutions are possible (longest edge up to 3840 pixels). This module supports continuous conversations. AI can make mistakes. Please verify important information.