Event
AGIBOT publishes Act2Goal goal-conditioned policy framework
Key points
- AGIBOT Finch published Act2Goal as a unified autoregressive framework that jointly pre-trains vision-language reasoning and robotic action generation.
- The official research page highlights strong simulated benchmark gains, especially on medium- and long-horizon goal-conditioned tasks using its multi-scale temporal hashing strategy.
- It matters because this is a first-party research release aimed at turning world-model planning into reusable goal-conditioned robot control rather than a one-off demo.
Company context
Develops humanoid robotic systems focused on general-purpose task execution, combining physical platforms with AI-driven control systems. The company is building a range of humanoid architectures targeting both industrial and service-oriented environments.
Context
- Company
- AgiBot
- Segment
- Humanoid
- Event type
- Research Publication
- Geography
- Shanghai · China