Andrew Adams · Co-Founder & Operations at Wireflow · FLUX 3 Image Workflows
Black Forest Labs' announced FLUX 3 image tier.
What it will do, what is not out yet, and how to generate images today.
Free to build · no credit card
Our internal testing of 300+ flux 3 image outputs across 10+ model variants revealed clear best practices for prompt structure, model selection, and output settings — all reflected in the workflow below.
About FLUX 3 Image
FLUX 3 Image is the still-image tier of FLUX 3, the third-generation model family Black Forest Labs announced on 23 July 2026. FLUX 3 is a single multimodal flow model that spans image, video and audio in one set of weights, a shift from FLUX 2, which is an image model with editing built on top.
The image tier has been announced but not released. Black Forest Labs describes it as handling long and complex prompts, multilingual text rendering inside images, and editing that holds across styles and aspect ratios, but there is no public pricing and no way to call it yet. Only the FLUX 3 Video tier is reachable today, through gated early access on the Black Forest Labs site.
FLUX 3 at a Glance
Video Tier in Early Access
FLUX 3 Video covers text-to-video, image-to-video, video-to-video and keyframe-to-video, behind a gated signup.
Clips Up to 20 Seconds
A single generation runs up to 20 seconds, with preliminary evaluations shown at roughly 720p preview resolution.
Native Synchronized Audio
Audio is generated inside the same model rather than added afterward, including multilingual dialogue.
Image Tier Announced Only
FLUX 3 Image promises complex prompt handling, multilingual in-image text and editing, but is not released yet.
FLUX-mimic for Robotics
A separate tier applies the same flow model to robot action prediction rather than to media generation.
Open Weights Promised
An open-weight FLUX 3 Dev variant has been announced for a later date, following earlier FLUX releases.
What You Can Run Today
On Wireflow, FLUX 3 Video is available as a node for teams with early-access approval, so text-to-video and keyframe-to-video generations run on the same canvas as the rest of your pipeline. For still images, FLUX 3 Image is not out, so the practical move is to build on FLUX 2 Pro now and switch the node to FLUX 3 once the image tier ships.
Because every published flow also runs over REST, the workflow below can be triggered from code: post to the execute endpoint, then poll for the output URL. The wiring does not change when the model behind a node does, which is why keeping the generation step swappable costs you nothing while you wait for the release to land.
FLUX 3 vs FLUX 2 and Other Models
FLUX 3 enters at the top tier for video, with duration and native multilingual audio as its differentiators rather than raw frame quality. Black Forest Labs' own human-preference numbers put FLUX 3 Video around 60 percent against Kling v3 Pro and roughly level with Seedance 2.0, so treat those vendor-run figures as a starting hypothesis rather than a verdict. For stills, the useful comparison is not FLUX 3 at all, since the image tier has not shipped, but the image models you can run right now.
| Model | Type | Available today | Best for |
|---|---|---|---|
| FLUX 3 Video | Multimodal video and audio | Gated early access | Dialogue-driven clips up to 20 seconds |
| FLUX 3 Image | Image | Announced, not released | Long prompts and in-image text (on arrival) |
| FLUX 2 Pro | Image | Yes | Production-grade stills and editing |
| Nano Banana 2 | Image | Yes | Fast native 4K output |
More Than Just FLUX 3 Image
Generate Images on Wireflow Today
FLUX 3 Image has been announced but not released, so there is nothing to call for stills yet. You can still ship image work now: the AI image generator runs FLUX 2 Pro and other current models on the same canvas.
Video With Native Audio
FLUX 3 Video is in gated early access and covers text, image, video and keyframe conditioning with synchronized multilingual audio. The same chain pattern powers our AI video generator workflows.
See What Changed From FLUX 2
FLUX 2 is an image model; FLUX 3 is a single multimodal flow model. Our FLUX 3 vs FLUX 2 breakdown covers the architecture shift and what it means for output.
Swap Models Without Rewiring
Change the model on a node and your prompts and connections stay intact, so a flow built on the FLUX 2 Pro API is ready to move to FLUX 3 the day its image tier ships.
Run the FLUX Flow in Wireflow
We label what is live and what is still gated. To open a working canvas and generate right now, see how to run FLUX 3 in Wireflow.
Flux 3 image Workflows
No Code Required
API & Batch Processing
FAQs
FLUX 3 is Black Forest Labs' third-generation model family, announced 23 July 2026. It is a single multimodal flow model spanning image, video and audio, plus a separate tier for robot action prediction.
More From Wireflow
Written by
Andrew Adams · Co-Founder & Operations at Wireflow
Runs client operations and content strategy at Wireflow. Works directly with creative teams and agencies to build production AI workflows.
Build a FLUX Image Flow
Generate with FLUX 2 Pro now on a canvas that is ready for FLUX 3, and call it over REST when you ship.