AIモデル · flux

FLUX.2 Flex

High-quality image generation and editing with up to 10 references.

10 CR / RUN · flux · 306.6K RUNS

概要

flux-2-flex is built for high-quality image generation and editing, giving you fine control over how images are created and refined. It supports prompts, adjustable dimensions, guidance, steps, and seed-based reproducibility.

A standout feature is support for up to ten reference images, making it a strong fit for workflows that need visual consistency across edits, variations, or style-guided outputs. You can also tune aspect ratio, resolution, output format, quality, safety tolerance, and prompt upsampling to match your project needs.

強み

01High-quality generation and editing in one model
02Supports up to 10 reference images
03Flexible control over size, aspect ratio, and resolution
04Seed support for repeatable results
05Adjustable guidance, steps, and output quality

プロンプト例

Create a cinematic portrait of a futuristic explorer at golden hour, ultra-detailed, natural skin texture, dramatic lighting, 4:5 aspect ratio

このプロンプトを試す

Using the reference images, generate a polished product hero shot with clean studio lighting, soft reflections, and a premium editorial look

このプロンプトを試す

Edit this scene into a moody rainy-night city street with neon reflections, realistic atmosphere, and consistent composition

このプロンプトを試す

Design a fantasy book cover featuring an ancient floating temple above the clouds, epic scale, rich color, high detail

このプロンプトを試す

活用例

Reference-guided creative work

Use up to ten reference images to steer style, composition, or visual consistency. This is useful when you want outputs to stay aligned with an existing look.

Image editing and refinement

Go beyond text-to-image generation by editing images with prompt-based direction. It's a practical option for transforming scenes, mood, or visual detail.

Marketing and brand visuals

Create polished campaign images, product scenes, or editorial-style assets with control over format, quality, and composition. Flexible output settings help tailor results to different channels.

Iterative concept development

Adjust seed, steps, guidance, and dimensions to explore variations systematically. This makes it easier to compare ideas and refine a concept over multiple runs.

よくある質問