Black Forest Labs, a new AI venture from Germany, has debuted its first AI image generation tool, called Flux. The tool is receiving recognition for its advanced capabilities beyond what Midjourney offers. Flux is compatible with high-performance laptops and aims to democratize sophisticated AI image creation.
Midjourney is currently considered state-of-the-art in the field of AI-generated imagery. It stands out for its ability to produce high-quality, detailed, and artistically appealing images from text prompts.
Different Versions for Various Needs
Flux comes in three distinct versions: Pro, Dev, and Schnell. The Pro version targets commercial users, with enterprises like FreePik already embedding it. Dev is aimed at developers, and Schnell, the swiftest of the variants, is optimized for speed. These models can be employed on high-end laptops, particularly those with an Nvidia RTX 4090 GPU, making them suitable for both professional and personal use.
Today we release the FLUX.1 suite of models that push the frontiers of text-to-image synthesis. read more at https://t.co/49zTUK8Q5V pic.twitter.com/hmcKRIlizn
— Black Forest Labs (@bfl_ml) August 1, 2024
Early feedback indicates that while Flux excels in rendering human figures, it still lags behind Midjourney v6.1 in skin texture quality. Nevertheless, its open-source framework and compatibility with widely available hardware make it a compelling choice for a diverse user base.
Development and Team
Flux was developed by a team of engineers previously associated with Stability AI, notable for their work on Stable Diffusion. Key figures include Andreas Blattmann, Patrick Esser, and Robin Rombach, all of whom were original co-creators of the Stable Diffusion models.
Stable Diffusion became popular due to its open-source nature, allowing widespread access and customization, combined with its ability to generate high-quality images efficiently.
Installation and local running of Flux are facilitated through the Pinokio launcher, a feature within the Pinokio AI Browser, which allows users to install and run applications with a single click. For users who cannot run the model locally, platforms like NightCafe, Poe, Based Labs, Hugging Face, and Fal.ai offer Flux access. FreePik also plans to incorporate Flux into its system soon.
The open-source approach and multi-platform support significantly widen Flux's user base. NightCafe has already added Flux to its offerings, enabling users to juxtapose its outputs with models like Ideogram and Stable Diffusion. Poe integrates it through a conversational interface, and it's also featured on Hugging Face and other AI repositories.
Black Forest Labs is venturing into the text-to-video domain with a future open-source model. Their ambition is to pioneer advancements in deep learning models for media, enhancing both creative possibilities and efficiency.
Company Vision
Black Forest Labs aspires to make leading generative AI tools widely accessible. “Our goal is to develop advanced deep learning tools for media, pushing creative, efficient, and diverse boundaries,” the company blog states. By distributing their models broadly, they aim to educate the public and ensure trust in AI's safety.
