VLA models
RFM-1
United States
Early commercial robotics foundation model for warehouse manipulation, including a learned physics world model that predicts video outcomes of candidate actions. In August 2024 Amazon hired Covariant's founders and about a quarter of its staff and took a non-exclusive license to its models.
OrganizationCovariant
CountryUS
Release2024-03
Params (B)8
Open weightsno
Architecture8B any-to-any multimodal transformer performing autoregressive next-token prediction across text, images, video, robot actions and numerical sensor readings
Licensenon-exclusive license to Amazon to use Covariant’s robotic foundation models
Modalitiesvision, language, action, video, sensor
Embodimentsmanipulator
Training dataInternet data plus Covariant's proprietary multimodal warehouse production data (Covariant Brain fleet: picking, sortation, induction, depalletization since 2017)
Sources: Covariant (2024-03-11)IEEE Spectrum (2024-03-11)TechCrunch (2024-08-31)
← Back to comparator · VLA models