FLUX.2 is the latest generation of Black Forest Labs models, optimized for higher win rates, better prompt following, and production-ready quality.

How is FLUX.2 different from FLUX.1?

FLUX.2 adds multi-reference editing, stronger text rendering, HEX color control, and JSON structured prompts, plus higher benchmark performance.

Is FLUX.2 free to try?

Yes. New users can start with free credits to generate images and explore the models.

Which FLUX.2 model should I choose?

Use Pro for the best quality, Flex for multi-reference editing and text, Dev for open-weight development, and Klein for lightweight edge use.

Can I use FLUX.2 commercially?

Yes. Pro and Flex support commercial use. Dev is open-weight and may have restrictions depending on licensing.

NEW - November 2025

FLUX.2: Frontier Visual Intelligence

Experience FLUX.2, the most capable AI image model from Black Forest Labs. Built for production workflows, it delivers higher quality, stronger prompt following, and reliable results at scale.

Powered by Mistral-3 integration and a 66.6% win rate in text-to-image benchmarks, FLUX.2 sets a new standard for creative teams and builders.

66.6% win rate

Mistral-3 VLM

2-6 cents per image

Mode

Select Model

Image Size

Your Prompt

Ready to Create

Choose a model, write your prompt, and let FLUX.2 bring your vision to life.

FLUX.2 Model Family

Choose the model that matches your workflow. All FLUX.2 variants share core intelligence with different trade-offs.

Best Quality

FLUX.2 Pro

Zero-config, production-ready quality for commercial generation.

Try model

Flexible Editing

FLUX.2 Flex

Multi-reference editing, strong text control, and adjustable parameters.

Try model

Open Weights

FLUX.2 Dev

Open-weight model for developers and rapid prototyping.

Try model

Coming Soon

FLUX.2 Klein

Lightweight model for edge, mobile, and fast deployment.

Why FLUX.2?

New capabilities that raise visual quality, consistency, and control.

Mistral-3 Integration

24B vision-language backbone for deeper prompt understanding.

Benchmark Leader

66.6% T2I win rate, 59.8% single-edit, 63.6% multi-edit.

Better Economics

ELO ~1030-1050 at just 2-6 cents per image.

World Knowledge

Accurate cultural, historical, and technical representations.

Prompt Following

Faithful style adherence with precise composition control.

FLUX.2 vs FLUX.1: What Changed?

A quick snapshot of the biggest capability shifts.

Feature

FLUX.1

FLUX.2

Text-to-Image

~50% win rate

66.6% win rate

Single Reference Edit

41.2% win rate

59.8% win rate

Multi-Reference Edit

~35% win rate

63.6% win rate

VLM Backend

None

Mistral-3 (24B)

Cost per Image

4-8 cents

2-6 cents

Open Source

Dev only

Dev + Klein (Apache)

FLUX.2 Capabilities

A preview of the visual range and control you can expect.

Magazine Cover Typography

Accurate text rendering with premium layout control.

Night Car Scene Lighting

Cinematic lighting and reflective surfaces.

Character Consistency

Reliable identity and styling across references.

Style Transfer

Blend multiple aesthetics with fidelity.

Product Shot with HEX Color

Precise color control for brand palettes.

Multi-Reference Composition

Combine product, scene, and lighting cues.

Looking for FLUX.1 Models?

Explore FLUX.1 Pro, Schnell, Dev, and Kontext for legacy workflows.

Frequently Asked Questions

Everything you need to know about FLUX.2.

Create with FLUX.2 on NanoWow

Start generating high-fidelity images with the latest FLUX models today.

View Pricing