FLUX 3: Black Forest Labs launches real world model

Loading video

Black Forest Labs launches FLUX 3 as a real world model, not just text-to-video or text-to-image. It unifies image, video, and audio in a multimodal flow-matching architecture to learn how the world works, supporting both content generation and embodied action prediction.

FLUX 3 Black Forest Labs 世界模型 world model 多模态 multimodal flow matching 具身AI embodied AI research Twitter

Category: research

Author: @shao__meng

Date: 2026-07-24T00:00:00

Duration: 46.142s

Source: https://x.com/shao__meng/status/2080499451576377388

EchoWM: A Drivable World Model, Not Just a Clip

EchoWM: A Drivable World Model, Not Just a Clip

@JinWeiyang18434 · 20 views · 2026-08-24

世界模型World ModelEchoWM

Build Manipulation Policies with NVIDIA Cosmos 3 Edge

Build Manipulation Policies with NVIDIA Cosmos 3 Edge

@NVIDIARobotics · 10 views · 2026-08-19

世界模型world model具身智能

Roblox launches its own world model

Roblox launches its own world model

@LeoKharon · 5 views · 2026-08-06

世界模型RobloxControlNet

Wayve Releases GAIA-4 World Model for Robotics Simulation

Wayve Releases GAIA-4 World Model for Robotics Simulation

@alexgkendall · 7 views · 2026-08-03

世界模型WayveGAIA-4

X-NavDP: Diffusion Navigation Policy Beyond Demonstrations

X-NavDP: Diffusion Navigation Policy Beyond Demonstrations

@robotsdigest · 9 views · 2026-08-01

导航navigation扩散策略

Why Reward-Based RL Is Extremely Sample Inefficient

Why Reward-Based RL Is Extremely Sample Inefficient

@cybernetic_lab · 10 views · 2026-07-31

强化学习reinforcement learningsample efficiency