FLUX 3: A Unified Multimodal Model for Real-World Visual Intelligence
We introduce FLUX 3, a new multimodal foundation model that jointly learns from images, videos, and audio within a unified architecture. By treating these modalities as projections of a single underlying reality, FLUX 3 captures physical laws and causal relationships better than isolated models. Early results in content creation and physical AI demonstrate its ability to perceive, predict, and act across diverse environments.
No single modality provides a complete description. Each is a projection of the same underlying reality, captured by different sensors, each of which loses some information in the process.
- user43928
I hope the open-weight versions will be SOTA.
> Over the next few weeks and months, we will make the following capabilities available
> Open-weight access to a multimodal backbone, for content creation (video, audio and image) and action prediction. (“FLUX 3 Dev”)
> We will also release more technical details on the underlying approach.
- thisisauserid
- Showed close to zero examples of people.
- Frivolous use of the term World Model.
- Claims 20 seconds of video, shows only jumpcuts.
Coming soon!
- jdthedisciple
It's incredible how negative and dismissive the comments in here are while here I am thinking the model actually looks impressively capable.
But then again I heard the downers have always been the first to leave their dung comments here so let's see...
- forgotusername6
Is anyone feeding models touch data? It seems the main thing we want the robots to do is touch things, but we are just feeding them audio/video/images. The model has to learn how to touch things despite never having touched anything before. Perhaps that's why they all look so hesitant when they touch things?
- AmbroseBierce
Amazing. This will be fundamental for the future for robots to distinguish the sound of the poors getting close to Besos/Musk's/Zuckerberg bunkers and quickly adapt to any new kind of attack by the masses, robots will quickly learn to adapt to the behavior of the attackers, quickly infer where they are grouped, their numbers and so for.
Of course there will be feuds from robots of different family groups but they will be minimal as it quickly becomes symmetrical robot conflict with high casualties as they learn too fast from each other, it's likely those will be avoided, it will be after all much easier to confront humans for any given resources.
Truly a pinnacle for technology, albeit perhaps not for mankind.
- Gecko4072
I thought the clips were real footage until they were dancing in a flooded room.
- make_it_sure
first AI thing coming from Europe that gives high hopes
- tormeh
These people are hiring in... Freiburg im Breisgau? Wonder how hiring is working out for them there.