SOURCE-REVIEWED MODEL PROFILE
Cosmos3-H-Dreams
An action-conditioned surgical world model that generates video from robot state and action inputs, rather than serving as a general chat assistant.
Metadata reviewed 2026-10-24 · Original creator: NVIDIA
What it does
- Mixture-of-Transformers architecture
- Kinematic action conditioning
- Distilled autoregressive video generation
Local fit
Specialized research deployment. Requires its dedicated research pipeline. It is not compatible with the launcher’s ordinary text-chat workflow and has not been tested locally by us.
Publisher/model configuration specification. Usable context depends on deployment resources.
Evidence and limits
Source-reviewed research profile; this exact checkpoint has not been benchmarked locally.
Research profile only
License and source
Custom NVIDIA license, including base-model conditions. Research coverage, not a bundled launcher model.
Read the original model card ↗
Reviewed source revision: fd87afe3fd97e9b352b09d9ada88c944f5621eb9