Black Forest Labs launches FLUX 3 Action, an open robotics AI model
Black Forest Labs is moving into robotics with FLUX 3 Action, an open world-action model that uses multi-camera video feeds to predict a robot’s next move and how its environment may change.
What FLUX 3 Action is built to do
Black Forest Labs is releasing FLUX 3 Action, an open AI model designed for robotics. The model builds on the multimodal FLUX 3 model, which was trained mainly on video as well as image and audio data.
As a world-action model, it takes multi-camera video feeds from a robot workspace and predicts what action an agent should take next. It also predicts how the environment will change as a result, making it relevant for robots that need to react to visual input.
Why the benchmark result matters
According to Black Forest Labs, FLUX 3 Action sets a record for success rate on the RoboLab-120 leaderboard while using just seven billion parameters. The company says it is less than half the size of the previous best open model and runs up to 3.95 times faster.
The key takeaway is efficiency: large reasoning models may plan well, but BFL says they can be too slow and bulky for robots. A smaller, faster model could be more practical for on-device deployment where response time and compute limits matter.
Where developers can watch next
The model weights are available on Hugging Face, according to the article. That makes FLUX 3 Action relevant for teams tracking open robotics models and comparing performance, speed, and deployment requirements.
BFL also points to possible uses beyond physical robots, including digital environments. Video games could serve as testing grounds for navigation, while fast-reacting agents that operate computers may follow later.