Introduction of FLUX 3 and FLUX-mimic Enhances Robotics and Video-Action Models

Jul 24, 2026

AI Summary

FLUX 3, a new multimodal foundation model, is now operational in robotics through a collaboration with mimic robotics, resulting in the FLUX-mimic video-action model. This model integrates video, audio, and action prediction, enabling robots to perform complex tasks in production environments, such as those at Audi, enhancing automation capabilities.

FLUX 3 is a multimodal foundation model that generates audio-visual content and predicts actions, developed in collaboration with mimic robotics.
The model has been tested in real-world applications, particularly in Audi's production lines, demonstrating its ability to handle complex tasks involving flexible parts.
FLUX 3's training involves significant computational resources, focusing on video prediction, which is critical for understanding physical interactions and dynamics.
The integration of action prediction into FLUX 3 did not compromise its existing capabilities, allowing it to maintain high performance in video generation tasks while adding new functionalities.
FLUX-mimic, built on the FLUX 3 backbone, is designed for general-purpose manipulation in industrial settings, leveraging learned world representations to improve task adaptability and efficiency.
The model has shown improved sample efficiency, requiring less demonstration data to achieve high success rates in robotic tasks compared to previous models.
Audi's partnership with mimic and Black Forest Labs aims to enhance production efficiency and expand flexible automation through the deployment of FLUX-mimic in their automated production network.
The FLUX-mimic system is optimized for rapid response times, achieving reaction times comparable to human visual processing, which is crucial for real-time operations in manufacturing.

video-action modelsflux 3mimicmachine learningai research

Introduction of FLUX 3 and FLUX-mimic Enhances Robotics and Video-Action Models

Related Stories

Leadership Changes at Google DeepMind Amidst Challenges

New AI Model Enhances Physical Simulations for Engineering Applications

Meta introduces Muse Glimmer AI model for personal use and privacy-focused applications

Meta to Open Source AI Models to Compete with OpenAI and Anthropic