A compact 450M-parameter VLA from Hugging Face, pretrained on community-collected LeRobot datasets and sized for fine-tuning and inference on consumer GPUs.
CogACT/CogACT-Base
Indexed from Hugging FaceLinked to its home on Hugging Face. FrontierOS indexes it; nothing is hosted or mirrored here.ModelVision-language-action7.6B paramsEst. 24 GB GPUEstimated minimum compute for inference: Desktop GPU with about 24 GB of memory. Not a benchmark.MITSingle arm
About
A VLA that uses a 7B vision-language model for high-level understanding and passes its output to a separate diffusion-transformer module that generates continuous action sequences.
Why it matters for physical AI
Its paper reports that splitting reasoning from a dedicated action module beats token-based VLAs like OpenVLA on simulated and real arm benchmarks, which makes it a useful architecture reference.
Summary written by FrontierOS in our own words. The upstream card and license govern use. Metadata checked 27 Sep 2026. Read the full card on Hugging Face
Files, weights and version history live on Hugging Face. FrontierOS indexes this repo; it does not host or mirror it.
hf download CogACT/CogACT-BaseUses the Hugging Face CLI (pip install -U huggingface_hub). This downloads the whole repo; add --include "<pattern>" to fetch only the files you need.
More vision-language-action
Browse allLeRobot's port of the Physical Intelligence pi0.5 base model, which is co-trained on web and multi-robot data to generalize better to unfamiliar homes and objects.
A 3B cross-embodiment robot foundation model from NVIDIA that combines a vision-language backbone with a diffusion-transformer action head and can be post-trained on a new robot with a modest amount of data.