Repository files navigation

AgiBot World Colosseo is a full-stack large-scale robot learning platform curated for advancing bimanual manipulation in scalable and intelligent embodied systems. It is accompanied by foundation models, benchmarks, and an ecosystem to democratize access to high-quality robot data for the academic community and the industry, paving the path towards the "ImageNet Moment" for Embodied AI.

We have released:

• GO-1: Our robotic foundation model pretrained on AgiBot World Dataset
• GO-1 Air: GO-1 model without Latent Planner, high-performanced and lightweighted
• Task Catalog: Reference sheet outlining the tasks in our dataset, including robot end-effector types, sample action-text descriptions and more
• AgiBot World Beta: Our complete dataset featuring 1,003,672 trajectories (~43.8T)
• AgiBot World Alpha: A curated subset of AgiBot World Beta, containing 92,214 trajectories (~8.5T)

• [2025/09/19] 🚀Our robotic foundation model GO-1 open-sourced.
• [2025/03/10] 📄Research Blog andTechnical Report released.
• [2025/03/01] Agibot World Beta released.
• [2025/01/03]Agibot World Alpha Sample Dataset released.
• [2024/12/30] 🤖 Agibot World Alpha released.

• AgiBot World Alpha
• AgiBot World Beta
  • ~1,000,000 trajectories of high-quality robot data
• AgiBot World Foundation Model: GO-1
  • GO-1 fine-tuning script
  • GO-1 Air pre-trained checkpoint
  • GO-1 pre-trained checkpoint
  • Examples of using GO-1 model
• 2025 AgiBot World Challenge

• 1 million+ trajectories from 100 robots.
• 100+ 1:1 replicated real-life scenarios across 5 target domains.
• Cutting-edge hardware: visual tactile sensors / 6-DoF Dexterous hand / mobile dual-arm robots
• Wide-spectrum versatile challenging tasks
• General robotic policy pretrained on AgiBot World

Contact-rich Manipulation	Long-horizon Planning	Multi-robot Collaboration
Fold Shirt (AgileX)	Fold Shirt (AgiBot G1)	Fold Shirt (Dual Franka)

• News📰
• TODO List 📅
• Key Features 🔑
• Table of Contents
• Getting started 🔥
  • Installation
  • How to Get Started with Our AgiBot World Data
    • Download Datasets
    • Visualize Datasets
  • How to Get Started with Our GO-1 Model
    • Requirements
    • Model Zoo
    • Fine-tuning on Your Own Dataset
    • Testing Your Model
    • More Examples
• License and Citation📄

1. Download our source code:

git clone https://github.com/OpenDriveLab/AgiBot-World.gitcd AgiBot-World

2. Create a new conda environment:

conda create -n go1 python=3.10 -yconda activate go1

3. Install dependencies:

This project is built onLeRobot (datasetv2.1, commit2b71789)
⚡️ Our environment has been tested withCUDA 12.4.

pip install -e.pip install --no-build-isolation flash-attn==2.4.2

If you encounter out of RAM issue while installingflash attention, you can set the environment variableMAX_JOBS to limit the number of parallel compilation jobs:

MAX_JOBS=4 pip install --no-build-isolation flash-attn==2.4.2

• [OPTION 1] Download data from ourOpenDataLab page.

pip install openxlab# install CLIopenxlab dataset get --dataset-repo OpenDriveLab/AgiBot-World# dataset download

• [OPTION 2] Download data from ourHuggingFace page.

huggingface-cli download --resume-download --repo-type dataset agibot-world/AgiBotWorld-Alpha --local-dir ./AgiBotWorld-Alpha

Convert the data toLeRobot Dataset format followingany4lerobot.

We adapt and extend the dataset visualization script fromLeRobot Project:

python scripts/visualize_dataset.py --task-id 390 --dataset-path /path/to/lerobot/format/dataset

It will openrerun.io and display the camera streams, robot states and actions, like this:

We strongly recommend full fine-tuning for the best performance. However, if GPU memory is limited, you can alternatively fine-tune only the Action Expert.

Here we provide an example of fine-tuning the GO-1 model on theLIBERO dataset. You can easily adapt it for your own data.

1. Prepare Data

We use the LeRobot dataset for our default dataset and dataloader. We provide a script for converting LIBERO to LeRobot format inevaluate/libero/convert_libero_data_to_lerobot.py.

Since TensorFlow is required to read theRLDS format, we recommend creating a separate conda environment to avoid package conflicts:

conda create -n libero_data python=3.10 -yconda activate libero_datapip install -e".[libero_data]"

Download the raw LIBERO dataset fromOpenVLA, then run the script to convert it into LeRobot dataset:

# Optional: Change the LeRobot home directoryexport HF_LEROBOT_HOME=/path/to/your/lerobotpython evaluate/libero/convert_libero_data_to_lerobot.py --data_dir /path/to/your/libero/data

2. Prepare Configs

We provide an example config for fine-tuning GO-1 on LIBERO ingo1/configs/go1_sft_libero.py.

Key sections in the config:

• DatasetArguments - path or repo for the LeRobot dataset.
• GOModelArguments - model settings: architecture (GO-1 Air or GO-1), action chunk size, diffusion scheduler, parameter freezing, etc.
• GOTrainingArguments - training hyper-parameters, seetransformers docs for more details.
• SpaceArguments - state/action dimensions, data keys in the LeRobot dataset, default language prompt, control frequency.

Seego1/configs/go1_base_cfg.py for all available config options.

3. Start Fine-tuning

Start fine-tuning with the following command, you can setup environment variables according to theshell.

RUNNAME=<YOUR_RUNNAME> bash go1/shell/train.sh /path/to/your/config

Checkpoints will be saved inexperiment/<YOUR_RUNNAME> and logs will be saved inexperiment/<YOUR_RUNNAME>/logs.

Notes:

• We also provide adebugging shell which can run on a single RTX4090. It also setDEBUG_MODE to true for faster init.
• We do not need to precompute the normalization statistics for the training data, as LeRobot will compute them when loading the dataset. The statistics will be saved toexperiment/<YOUR_RUNNAME>/dataset_stats.json.
• We set action chunk size and control frequency input as 30 in GO-1 pre-training, as our AgiBot World dataset is collected at 30Hz. We change them to 10 in LIBERO fine-tuning, as the LIBERO dataset is collected at 10Hz. You can change them accordingly in the config file.

Local Inference

After fine-tuning, you can test your model locally using an example script inevaluate/deploy.py. You can build aGO1Infer object to load the model and dataset statistics, then call theinference method to run inference:

importnumpyasnpfromevaluate.deployimportGO1Infermodel=GO1Infer(model_path="/path/to/your/checkpoint",data_stats_path="/path/to/your/dataset_stats.json")payload= {"top": ...,"right": ...,"left": ...,"instruction":"example instruction","state": ...,"ctrl_freqs":np.array([30]),}actions=model.inference(payload)

We also provide a script for open-loop evaluation with training data inevaluate/openloop_eval.py.

Remote Inference

Considering that 1. real robot may not have powerful GPUs, 2. different robots and simulation benchmarks often require different package dependencies, we also provide a policy server for GO-1. A client in another environment or another machine send observations to the server for remote inference.

Start the server and it will listen on portPORT and waits for observations:

python evaluate/deploy.py --model_path /path/to/your/checkpoint --data_stats_path /path/to/your/dataset_stats.json --port<PORT>

For the client, we provide aGO1Client class to send requests to the server and receive actions:

fromtypingimportDict,Anyimportjson_numpyimportnumpyasnpimportrequestsjson_numpy.patch()classGO1Client:def__init__(self,host:str,port:int):self.host=hostself.port=portdefpredict_action(self,payload:Dict[str,Any])->np.ndarray:response=requests.post(f"http://{self.host}:{self.port}/act",json=payload,headers={"Content-Type":"application/json"}      )ifresponse.status_code==200:result=response.json()action=np.array(result)returnactionelse:print(f"Request failed, status code:{response.status_code}")print(f"Error message:{response.text}")returnNone

We can then run the LIBERO evaluation script to query the server, see theLIBERO README for details.

We will provide more examples of fine-tuning and running inference with GO-1 models on real robots and simulation platforms.

Currently we have:

• Genie Studio: AgiBot G1 with out-of-the-box GO-1 model plus integrated data collection, fine-tuning, and deployment pipeline.
• AgileX: AgileX Cobot Magic (Aloha)
• LIBERO: LIBERO Simulation (Franka)
• RoboTwin: RoboTwin Simulation (Aloha)

All the data and code within this repo are underCC BY-NC-SA 4.0.

• Please consider citing our work if it helps your research.
• For the full authorship and detailed contributions, please refer tocontributions.
• In alphabetical order by surname:

@article{bu2025agibot_arxiv,title={Agibot world colosseo: A large-scale manipulation platform for scalable and intelligent embodied systems},author={Bu, Qingwen and Cai, Jisong and Chen, Li and Cui, Xiuqi and Ding, Yan and Feng, Siyuan and Gao, Shenyuan and He, Xindong and Huang, Xu and Jiang, Shu and others},journal={arXiv preprint arXiv:2503.06669},year={2025}}@inproceedings{bu2025agibot_iros,title={Agibot world colosseo: A large-scale manipulation platform for scalable and intelligent embodied systems},author={Bu, Qingwen and Cai, Jisong and Chen, Li and Cui, Xiuqi and Ding, Yan and Feng, Siyuan and He, Xindong and Huang, Xu and others},booktitle={2025 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},year={2025},organization={IEEE}}@article{shi2025diversity,title={Is Diversity All You Need for Scalable Robotic Manipulation?},author={Shi, Modi and Chen, Li and Chen, Jin and Lu, Yuxiang and Liu, Chiming and Ren, Guanghui and Luo, Ping and Huang, Di and Yao, Maoqing and Li, Hongyang},journal={arXiv preprint arXiv:2507.06219},year={2025}}

@misc{AgiBotWorldTeam2025agibot-world-colosseo,title        ={Introducing AgiBot World Colosseo: A Large-scale Manipulation Platform for Scalable and Intelligent Embodied Systems},author       ={Shi, Modi and Lu, Yuxiang and Wang, Huijie and Xie, Chengen and Bu, Qingwen},year         ={2025},month        ={March},howpublished ={\url{https://opendrivelab.com/AgiBot-World/}},note         ={Blog post},        }@misc{AgiBotWorldTeam2025open-sourcing-go1,title        ={Open-sourcing GO-1: The Bitter Lessons of Building VLA Systems at Scale},author       ={Shi, Modi and Lu, Yuxiang and Wang, Huijie and Yang, Shaoze},year         ={2025},month        ={September},howpublished ={\url{https://opendrivelab.com/OpenGO1/}},note         ={Blog post},        }