Light-O1-Preview
Light-O1: Scaling Whole-Body Intelligence with Human Action Pretraining
Light-O1-Preview is a reasoning text-to-action model: given an instruction in natural language it reasons about intent and constraints, then generates a sequence of whole-body humanoid actions. Reasoning text and actions are decoded in one autoregressive stream — action codes are ordinary rows of the embedding table rather than a task-specific prediction head — so the model's reading of an instruction can be inspected directly instead of inferred from the resulting motion. Actions are produced in a shared humanoid representation rather than for one specific robot, so a generation can be retargeted to other embodiments.
[!NOTE] Light-O1-Preview releases the text-to-action capability only. We are working towards the full Light-O1, which will add vision and manipulation. Stay tuned — thanks for your interest!
Inference, serving and the operator console: Light-O1 repository. Try it without any local setup in the Light-O1-Preview Playground.
Files
| Path | Content |
|---|---|
config.json, model.safetensors | model weights |
tokenizer.json, tokenizer_config.json, chat_template.jinja, processor_config.json, generation_config.json | tokenizer / processor |
human_action_138_v1.json | the action representation that must match training (field layout, 22-joint skeleton, frame rate, axis convention); read automatically by the inference code |
action_tokenizer/ | FSQ action-decoder bundle (codebook 65,536, four levels of 16, 20 fps) |
Usage
hf download LightOriginsHQ/Light-O1-Preview --local-dir ./Light-O1-Preview
git clone https://github.com/lightorigins/Light-O1 && cd Light-O1
uv sync --extra inference
# one-off generation on the command line
uv run --extra inference light-deploy --model /absolute/path/to/Light-O1-Preview \
--prompt "a person waves with the right hand" --thinking --output human_action.npy
# local WebUI: streamed reasoning and a 3D preview of the generated action
uv run --extra inference light-deploy-server \
--model-path /absolute/path/to/Light-O1-Preview --port 8090
GPU inference needs Linux x86-64, Python 3.11 and a CUDA 13 compatible environment. The HTTP API, the split GPU / Control host setup and the optional MuJoCo simulation example are documented in the code repository.
Model details
- Input: a natural-language prompt.
- Output: a reasoning trace followed by action codes, decoded to
human_action_138_v1— per-frame root displacement in the ground plane, pelvis height, yaw rate, 22 parent-local 6D joint rotations and two hand-open scalars, right-handed and Y-up at 20 fps.
Community
Questions, deployment notes and release news — join us on Discord, or scan to join the WeChat group:
License
Apache License 2.0 (see LICENSE). Required upstream notices are in
THIRD_PARTY_NOTICES.md and LICENSES/.