Reve opens API for its 4K image model
Reve has launched the Reve 2.1 API, which gives developers its native 4K image model and, for the first time outside Reve.com, the layout model that turns images into structured code that agents can edit.
Image-generation startup Reve announced the Reve 2.1 API, which makes its image model and its underlying layout model available to outside developers.
The company says the model generates images at native 4K x 4K resolution (16 megapixels) with no upscaling step, which it credits to a "novel rendering architecture." Reve calls it "the world's best 4K image model." The announcement includes no benchmarks or comparisons to back that claim.
The API comes in two tiers. The "Ready to use" endpoints handle standard image creation and editing, and Reve says a single Create endpoint covers most integrations. Reve describes them as offering consistent style and content references for brand and campaign work, along with strong typography and graphic design.
The second tier, "Full Layout Control," is labeled experimental. It splits layout creation, editing and rendering into separate endpoints. Reve says it bet two years ago on layouts as an intermediate representation for images: hierarchical regions whose shape, color and meaning can each be addressed and changed on their own. Because each step is separate, developers can re-run only the step that changed, render one layout as a batch of variations, or apply a single edit across many images. Reve also says per-region processing makes small text and complex scenes more accurate.
Reve is pitching the layout tier as "agent-native." It ships a skill.md file so coding agents can read an image's structure and edit individual elements instead of just requesting a finished render.
The post does not give pricing, rate limits or model details. Developers can start from Reve's API console, and enterprise customers are directed to sales.