Perceptron Inc releases SotA VLA model, unifying perception and control
code_star · x · 2026-08-20
Perceptron Inc has announced the training of a state-of-the-art VLA (Vision-Language-Action) model, with plans for an open-source release.
Key Insights:
- Unified Intelligence: The split between perception and control is seen as an artifact of field history, not physical intelligence. Their next model is a true embodied foundation model jointly trained as a VLM, ER, and VLA.
- Transfer Learning: The data and modeling efforts for their frontier VLM were found to have substantial transfer learning benefits for training the VLA.
- Upcoming Release: The team is sprinting towards an open-source release of the model and the accompanying paper.
Related event: Perceptron Releases SotA Embodied VLA Model, Plans to Open Source(2 posts)→
More from Embodied
- CMU's Andrea Bajcsy on what safety means for generalist robots, beyond collision avoidance — CSProfKGD · 2026-10-16
- Codex drives a physical CAN-bus motor, eyeing a fully Codex-built robot — OpenAIDevs · 2026-10-16
- Berkeley paper uses the Midas myth to rethink robotics beyond human imitation — begusgasper · 2026-10-16
- Team uses Claude Opus 5.5 to reproduce core claims of 10 recent robotics papers end to end — teortaxesTex · 2026-10-16
- Kinex: a harness that connects coding agents to any physical robot for scalable learning — daibond_alpha · 2026-10-16
- Gemini is a robotics outlier: dropping it lifts capability-robotics correlation from 0.20 to 0.96 — ycombinator · 2026-10-16