AI Models · flux

FLUX 3

Generate or edit images with prompts and up to 10 references.

10 CR / RUN · flux · 241.8K RUNS

FLUX 3

About

FLUX 3 from Black Forest Labs is a text-to-image and image-editing model built for flexible visual creation. Start from a written prompt or guide the result with reference images.

It supports up to 10 input images, giving you more control when refining style, composition, or subject details. With options for resolution, aspect ratio, output format, quality, grounding, and safety tolerance, it fits a wide range of image workflows.

Strengths

  • 01Creates images directly from text prompts
  • 02Edits images using up to 10 reference images
  • 03Offers control over resolution and aspect ratio
  • 04Supports output format and quality settings
  • 05Includes grounding and safety tolerance inputs
  • 06Widely used on Replicate

Example prompts

A cinematic portrait of a traveler in a rain-soaked neon city at night, dramatic lighting, detailed reflections

Try this prompt

Minimalist product photo of a matte black coffee grinder on a stone surface, soft studio lighting, clean background

Try this prompt

A cozy mountain cabin at sunrise, pine trees in the mist, warm window glow, natural color palette

Try this prompt

Edit these reference images into a polished fashion campaign look with consistent lighting, modern editorial styling, and a premium studio finish

Try this prompt

Use cases

Text-to-image concepting

Turn written ideas into original visuals for campaigns, moodboards, or creative exploration. It's a fast way to move from concept to image.

Reference-based image editing

Use up to 10 reference images to guide edits and keep outputs aligned with your visual direction. Helpful for refining style, subjects, or composition.

Marketing and content visuals

Create images sized for different channels with adjustable aspect ratio and resolution. Useful for social posts, landing pages, and promotional assets.

Controlled output workflows

Fine-tune results with settings for output format, quality, grounding, and safety tolerance. This makes it easier to adapt the model to different production needs.

FAQ

Related models