Set the visual direction
Choose Sunburst or Flare in the generator to explore the available variants.
GPT Image 2.5
Build a visual from a detailed brief, then refine it with image references. Explore campaign concepts, product compositions and editorial imagery.
Up to 4K output
5 quality tiers
Up to 16 references
Meet GPT Image 2.5
GPT Image 2.5 is from OpenAI. Explore these workflows and choose generation settings for your project.
Shape the workflow
Bring up to 16 images into an editing workflow for visual context.
Choose your settings
Select a quality tier and output size to suit a draft or a detailed final composition.
Put it to work.
Explore practical workflows and load a prompt to try your own direction.
Posters & campaigns
Develop campaign posters with a clear focal point, deliberate spacing and a considered palette. Explore the composition before finishing your final typography.
Product photography
Explore studio lighting, materials and lifestyle settings for your next product campaign. Start with a concept, then refine it with your own product references.
Social content
Build a visual idea for launches, everyday stories and short campaigns. Start with one strong image and carry the same direction into your next variation.
From an idea to a first result.
Describe the scene
Define the subject, setting, light and composition. For video, add action and camera direction.
Choose your settings
Confirm the model in the generator and choose from its available size, duration or quality options.
Generate and refine
Review the first result, then adjust the prompt or add reference material to develop the idea.
A few things to know
Useful answers before you start creating.
What can I make with GPT Image 2.5?
Build a visual from a detailed brief, then refine it with image references. Explore campaign concepts, product compositions and editorial imagery.
Which settings are available here?
Up to 4K output; 5 quality tiers; Up to 16 references. Available options depend on the variant selected in the generator.
Are these images outputs from this model?
The images come from the corresponding public WaveSpeed model pages, linked below the image. They are not viddeep benchmarks or output guarantees. Results depend on the prompt, references and settings.
What do I need to start?
You can explore the tool and settings first. Submitting a generation requires sign-in and sufficient credits; the generator shows the cost before you submit.
Your next idea starts here.
Bring your brief into the studio.