Flux Image Models Are Now Accessible On Freepik
Blackforest Lab's suite of Flux image models is now available on Freepik.
In recent weeks, Flux.1, the latest image model from Blackforest Labs, has been going viral in the AI community. This open-weight model has been getting a lot of praise from both casual users and startups for creating images that are as impressive as those from Midjourney V6.1.
The question now is, how do you get access to Flux?
One of the first on the market to offer the full power of the Flux model in an easy-to-use product is Freepik. But before I get into the specifics of how Flux works on Freepik, it’s important to understand what this new image model brings to the table.
What is Flux?
Flux.1 is a new state-of-the-art (SOTA) suite of text-to-image models that set a new standard in image detail, prompt accuracy, style variety, and scene complexity for text-to-image generation.
It comes in three variants:
Flux.1 Pro: This model delivers next-level performance in image generation, excelling in prompt adherence, visual quality, image detail, and output diversity.
Flux.1 Dev: A guidance-distilled model tailored for non-commercial use, Flux.1 Dev is derived from Flux.1 Pro, offering comparable quality and prompt accuracy while being more efficient than typical models of its size.
Flux.1 Schnell: The fastest model in the suite, designed for local development and personal use, is available under an Apache 2.0 license.
All public Flux.1 models feature a blend of multimodal and parallel diffusion transformer blocks, boasting 12 billion parameters. These models outperform previous diffusion models by leveraging flow matching — a straightforward method for training generative models that incorporates diffusion.
Moreover, the models achieve better performance and hardware efficiency through the use of rotary positional embeddings and parallel attention layers.
Flux on Freepik
Freepik, already known as a massive repository of high-quality stock photos, videos, and vector illustrations, has taken a step forward by integrating Flux models into its AI tools. If you’re familiar with Freepik’s offerings, you’ll know they’ve been expanding their range of AI tools, including an AI image generator and mockup generator, as well as an AI image upscaler.
Recently, Flux models were added to Freepik’s lineup of supported image models within their AI image generator tools.
Flux Fast: The current model is in production. It excels at prompt adherence and generating text within images. It’s close to Midjourney in quality but not quite there yet.
Flux: This variant, on par with Midjourney, will be available under the name Flux.
Flux Realism: This is the top choice for photorealism, surpassing even Midjourney in creating lifelike photos.
To try it out, go to Freepik’s Pikaso suite of AI tools and select the “AI Image Generator.”
The image generator tool looks like this:
Under the model tab, select your preferred Flux mode. It’s not clear which image model is underneath each mode. My guess is that the model powering Flux fast is Schnell, while Flux Realism is the Pro model.
In the example section, I’ll demonstrate the image differences using a similar prompt. Also, take note that if you set the mode to Flux Realism, you won’t be able to change the style — it will stay on default.
Example Image Generation
There are two ways to generate an image on Freepik. You can either start with a text description or use an existing image as a reference. Let me give you an example that showcases the realism capability of the Flux model.
Keep reading with a 7-day free trial
Subscribe to Generative AI Publication to keep reading this post and get 7 days of free access to the full post archives.