What FLUX models are available?

Last updated: August 26, 2026

FLUX Models Overview

Black Forest Labs offers three model families: FLUX 3 (video with synchronized audio), FLUX.2 (image generation and editing), and FLUX.1 (including Kontext and FLUX1.1 variants).

FLUX 3: Video Generation

One multimodal model that generates video with synchronized audio: speech with lipsync, sound effects, and ambience are created together with the frames.

  • Text to Video, Image to Video, and Video Continuation on a single endpoint (flux-3-video)

  • Clips up to 20 seconds long, at hd or fhd resolution, 24 fps

  • Draft mode for fast, cheaper previews you can later render at full quality

  • Priced per second of video: full renders from $0.17/s, drafts from $0.06/s

FLUX.2: Latest Generation

Our most advanced image generation and editing models. All FLUX.2 models support resolutions up to 4MP, multi-reference editing, precise hex color control, and reliable text rendering.

FLUX.2 [klein]: Fastest

Fastest inference for real-time applications. Available in two sizes:

Variant

Parameters

License

Price (API)

[klein] 4B

4 Billion

Apache 2.0 (fully open)

from $0.014

[klein] 9B

9 Billion

FLUX Non-Commercial

from $0.015

  • Runs on consumer GPUs

  • Unified generation and editing architecture

  • Up to 4 reference images

  • Available on Hugging Face for local deployment

FLUX.2 [pro]: Production Ready

Optimized for high-volume production workflows. Generates images in under 10 seconds.

  • Best balance of quality, speed, and cost

  • Up to 8 reference images (API) / 10 (Playground)

  • Text-to-image: from $0.03/MP | Editing: from $0.045/MP

FLUX.2 [max]: Highest Quality

Maximum quality with exclusive grounding search, performs web searches to incorporate real-time information into generated images.

  • Highest editing consistency and vast world knowledge

  • Up to 8 reference images (API) / 10 (Playground)

  • Text-to-image and editing: from $0.07/MP

FLUX.2 [flex]: Fine-Grained Control

Maximum control with adjustable inference steps (up to 50) and guidance scale. Best text rendering for typography-heavy use cases.

  • Adjustable steps and guidance for precise control

  • Text-to-image: $0.05/MP | Editing: $0.10/MP

FLUX.2 [dev]: Open Weights

Free for non-commercial use. Run locally on your own infrastructure.

FLUX.1 Kontext: Context-Aware Editing

Combines text-to-image generation with advanced image editing in a single model. Edit specific parts of images, maintain character consistency, and transform styles with simple text prompts.

Model

Best For

Price

Kontext [max]

Highest quality, industry-leading typography

$0.08/image

Kontext [pro]

Fast production (5-6 sec), unified editing & generation

$0.04/image

Kontext [dev]

Open weights, editing mode only, non-commercial

Free

FLUX1.1: Text-to-Image

Model

Description

Price

FLUX1.1 [pro]

Fast, reliable text-to-image generation

$0.04/image

FLUX1.1 [pro] Ultra

Ultra high-resolution up to 4MP

$0.06/image

Tools

Model

Description

Price

FLUX.1 Fill [pro]

Inpainting and outpainting with text prompts and masks

$0.05/image

Which Model Should I Use?

  • Real-time generation? → FLUX.2 [klein]

  • Production workflows? → FLUX.2 [pro]

  • Highest quality? → FLUX.2 [max]

  • Typography and fine control? → FLUX.2 [flex]

  • Image editing with text prompts? → FLUX.1 Kontext [pro] or FLUX.2 with references

  • Running locally? → FLUX.2 [klein] 4B (Apache 2.0) or FLUX.1 Kontext [dev]

For full documentation, visit docs.bfl.ai.