← Embodied AI LLMs
AgiBot GO-1
Candidate (候选)智元机器人· VLA foundation models
Generalist embodied foundation model on the ViLLA architecture.
Capabilities
- First generalist embodied foundation model built on the ViLLA (Vision-Language-Latent-Action) architecture.
- Reusable latent-action planning across a wide range of manipulation tasks.
- Open weights plus a full training / data framework released to the community.
Specifications
License
Fully open source
Data residency
Weights downloadable / on-device
Overseas
Partial (units exported)
International API
No
Vertical
具身智能 · Embodied AI
Status
Candidate (editorial shortlist)
More Embodied AI LLMs
Want to bring this model into production?
Run a compliance and deployment assessment, and book vendor intro & PoC.