Science & Tech · AI · 2 hrs ago
BeingBeyond model uses 10,000 hours of video to train humanoid robots
BeingBeyond has developed Being-M0.7, a model designed to help humanoid robots walk and grasp.
It uses 10,000 hours of human video as training data.
Before acting, the model predicts what it will see and how its whole body should move.
In the SIMPLE simulation, it achieved a score of 71.1%, leading the results there.
On a Unitree G1 robot, it matched the strongest baseline.
These tests show how the model performed in simulation and on a robot, but no further plans or results are given.
BeingBeyond’s Being-M0.7 uses 10,000 hours of human video to train humanoid robots.
The model predicts future visual latents and whole-body motion before acting.
It led the SIMPLE simulation with a score of 71.1%.
On a Unitree G1 robot, it tied the strongest baseline.
- Who
- BeingBeyond developed the Being-M0.7 model.
- What
- The model uses 10,000 hours of human video to teach humanoids to walk and grasp.
- When
- Published on 2026-10-12.
- Where
- The model was evaluated in SIMPLE simulation and on a Unitree G1 robot.
- Why
- To teach humanoid robots to walk and grasp.
This story does not have two clearly opposing sides.
No direct quotes in the coverage so far.
This story does not have a timeline yet.
- Model
- Being-M0.7
- Training video
- 10,000 hours of human video
- SIMPLE simulation
- 71.1%
- Robot evaluation
- Unitree G1
Sources
Related news
Anthropic AI submits false murder tip to Philadelphia police
Nadella calls for human-controlled emergency brake on advanced AI
Three-quarters of online stores score poorly on AI shopping readiness
TikTok CEO predicts AI-driven scientific breakthroughs within three years
Microsoft expands Copilot Notebooks with productivity and automation upgrades
Chinese factories are adopting AI agents and taking the approach abroad
Pharma companies invest heavily as AI reshapes drug research
European rules slow rollout of remotely controlled river ships
OpenAI’s ChatGPT for Teens Study Mode Still Does Homework
Anthropic bars repeated cruelty toward its AI models
OpenAI fires three safety researchers after dispute over sensitive information
