DreamLayer AI is an image generation and editing platform. You describe the image you need, add a reference when identity or style should carry forward, and refine the result over several turns, in the browser or through an API, CLI, and MCP server. This leaderboard comes from DreamLayer Eval, its separate open-source benchmarking project.
DreamLayer AI routes each request to one of several image models depending on the task. The models on this page were benchmarked by DreamLayer Eval in September 2025; the product's model choices are maintained separately and are not tied to this snapshot.
A single-model tool uses one model for every request. DreamLayer AI chooses among several image models per task, carries a reference identity forward across edits, and offers background removal and upscaling in the same conversation.
Yes. Upload the image as a reference and describe the change you want. DreamLayer AI applies the edit and keeps the rest of the image, and you can refine the result over further turns.
No. DreamLayer AI is a hosted product. It runs in the browser and through its API, CLI, and MCP server, and it manages model access for you. DreamLayer Eval, the open-source benchmarking project, is the part you run locally if you want to reproduce benchmarks.
One credit per finished image, across the browser, API, CLI, and MCP server. Credits do not expire, and an unfinished request restores the reserved credit. Details are on the pricing page.
DreamLayer Eval is benchmarking infrastructure for image and video diffusion models. It automates prompts, seeds, configs, metric scoring, and reproducible run logging so researchers can compare models consistently. The code is available on GitHub.
Built-in evaluation metrics include CLIP Score, FID, precision, recall, F1, LPIPS, SSIM, PSNR, and temporal consistency for video, all logged automatically with the prompts, seeds, and configs that produced them.
Yes. Every run is logged with its prompts, seeds, configs, outputs, and metric scores, and can be exported as CSV, JSON, or a complete benchmark bundle, so results stay traceable and repeatable.
DreamLayer Eval measures how models perform under fixed prompts, seeds, and configs, and the leaderboard on this page is its September 2025 result. DreamLayer AI is the separate hosted product. It does not depend on DreamLayer Eval, and its model choices are maintained on their own.