Config releases tech preview for bimanual robot data and foundation-model platform
Evidence notes
- Config published a technical preview describing its workflow for adapting foundation models to new bimanual robot tasks.
- The company said it had collected about 100,000 hours of human action data and was collecting around 20,000 hours per month.
- The preview described CFG-1 as a first-generation foundation model trained for bimanual manipulation.
Company context
Config builds robotics data infrastructure and foundation-model technology for bimanual manipulation. Its platform covers the workflow from task and gripper definition through human and teleoperation data acquisition, data validation, action estimation, model training and evaluation.