AIモデル · flux

FLUX Redux Dev

Create fresh image variations while keeping the core of your original.

10 CR / RUN · flux · 338.7K RUNS

概要

flux-redux-dev is an open-weight image variation model built for remixing an existing image into new outputs. It is designed to preserve key elements from your source while exploring different looks, compositions, and refinements.

With controls for guidance, aspect ratio, megapixels, inference steps, seed, and multiple outputs, it gives you a flexible way to generate alternate versions for creative exploration, iteration, and visual development.

強み

01Built specifically for image variation workflows
02Preserves key elements from the original image
03Open-weight model from Black Forest Labs
04Supports aspect ratio and megapixel controls
05Can generate multiple outputs in one run
06Seed control helps with repeatable experiments

プロンプト例

Create three polished variations of this portrait that keep the subject and overall identity intact, but explore softer cinematic lighting and a more premium editorial feel.

このプロンプトを試す

Remix this product image into clean studio-style variations while preserving the product shape, colors, and key design details.

このプロンプトを試す

Generate alternate versions of this character image that keep the character recognizable, but vary the framing, mood, and background atmosphere.

このプロンプトを試す

Create wide-format variations of this scene that maintain the main composition elements while adding richer environmental detail and depth.

このプロンプトを試す

活用例

Creative iteration

Generate multiple directions from a single source image without starting over. It's useful when you want fresh options that still feel connected to the original.

Style exploration

Test different visual treatments while preserving important content. This makes it easier to compare moods, lighting, and presentation styles from the same base image.

Format adaptation

Use aspect ratio controls to create new versions suited to different layouts. You can adapt a source image for wider or alternate compositions while keeping its key elements.