| Booster T1 | Velocity tracking with a gait clock (forward, lateral, yaw rate) | booster_gym (Isaac Gym) PPO, MLP 47-256-128-128-12, ELU | Apache-2.0 | 47 obs → 12 actions @ 50 Hz | runs in tab ▸ | github.com/BoosterRobotics ↗ |
| Unitree G1 | Velocity tracking on flat ground (forward, lateral, yaw rate) | legged_gym (Isaac Gym) + rsl_rl PPO, ActorCriticRecurrent (LSTM 64, MLP 32, ELU) | BSD-3-Clause | 47 obs → 12 actions @ 50 Hz · LSTM | runs in tab ▸ | github.com/unitreerobotics ↗ |
| Unitree H1 | Velocity tracking on flat ground (forward, lateral, yaw rate) | legged_gym (Isaac Gym) + rsl_rl PPO, ActorCriticRecurrent (LSTM 64, MLP 32, ELU) | BSD-3-Clause | 41 obs → 10 actions @ 50 Hz · LSTM | runs in tab ▸ | github.com/unitreerobotics ↗ |
| Aero Hand Open | in-hand cube rotation about Z (60 mm, 117 g cube) | MuJoCo Playground PPO, deployed via ROS 2 (rl_z_rotation_deploy) | not stated | in source | recipe | github.com/Chestnut-Robotics ↗ |
| AgiBot Lingxi X1 | walk, legs only (upper body held still), joystick velocity commands | ONNX Runtime in AimRT rl_control_module, 1 kHz control; trained with agibot_x1_train (Isaac Gym, humanoid-gym style PPO) | not stated | in source | weights | github.com/AgibotTech ↗ |
| AgiBot Lingxi X1 | walk with shoulder pitch following the gait (walk_leg_arm mode) | ONNX Runtime in AimRT rl_control_module | not stated | in source | weights | github.com/AgibotTech ↗ |
| AgiBot Lingxi X1 | x1_dh_stand training recipe (standing/walking, 12 leg actions) | Isaac Gym Preview 4 + rsl_rl-style PPO (legged_gym / humanoid-gym lineage) | not stated | in source | recipe | github.com/AgibotTech ↗ |
| ALOHA | Bimanual cube transfer (simulation, human demos) | LeRobot ACT | not stated | in source | weights | huggingface.co/lerobot ↗ |
| ALOHA | Bimanual peg insertion (simulation, human demos) | LeRobot ACT | not stated | in source | weights | huggingface.co/lerobot ↗ |
| Asimov 1 | Full-body velocity locomotion (23-DOF, neck locked) | Isaac Lab or MJLab templates (per docs); MuJoCo/URDF models in repo | not stated | in source | recipe | docs.menlo.ai/asimov ↗ |
| Asimov 1 | Sim model for training | MuJoCo 3.x | not stated | in source | recipe | github.com/menloresearch ↗ |
| Asimov 1 | Asimov v0 12-DOF biped velocity tracking (legs-only predecessor) | mjlab (MuJoCo Warp) RL | not stated | in source | recipe | github.com/menloresearch ↗ |
| Berkeley Humanoid Lite | Velocity-Berkeley-Humanoid-Lite-v0 (22-DoF full humanoid velocity tracking) | Isaac Lab 2.1.0 / Isaac Sim 4.5.0 + rsl_rl PPO; ONNX deploy (policy_dt 0.04 s = 25 Hz) | not stated | in source | weights | github.com/HybridRobotics ↗ |
| Berkeley Humanoid Lite | Velocity-Berkeley-Humanoid-Lite-Biped-v0 (12-DoF legs only) | Isaac Lab + rsl_rl PPO; ONNX (policy_dt 0.02 s = 50 Hz); also policy_biped_25hz_a/b.onnx | not stated | in source | weights | github.com/HybridRobotics ↗ |
| Berkeley Humanoid Lite | Humanoid legs-only policy | Isaac Lab + rsl_rl; ONNX | not stated | in source | weights | github.com/HybridRobotics ↗ |
| Berkeley Humanoid Lite | Video demo policy | Isaac Lab + rsl_rl; ONNX | not stated | in source | weights | github.com/HybridRobotics ↗ |
| Berkeley Humanoid Lite | Sim2sim validation in MuJoCo | MuJoCo | not stated | in source | recipe | berkeley-humanoid-lite.gitbook ↗ |
| Berkeley Humanoid Lite | VR teleoperation of arms (SteamVR bridge) | SteamVR-Bridge (Windows) + IK solver | not stated | in source | recipe | github.com/ucb-bar ↗ |
| Dropbear | walk, flat ground (v0.1.0, validated in sim) | Isaac Lab + RSL-RL PPO | not stated | in source | weights | github.com/Hyperspawn ↗ |
| Dropbear | walk over low obstacles (v0.2.0, experimental) | Isaac Lab + RSL-RL PPO | not stated | in source | weights | github.com/Hyperspawn ↗ |
| Duke Humanoid v1 | walk, velocity tracking (baseline) | Isaac Gym + rl_games (legged_env) | not stated | in source | weights | github.com/generalroboticslab ↗ |
| Duke Humanoid v1 | walk, energy-efficient passive-dynamics policy | Isaac Gym + rl_games (legged_env) | not stated | in source | weights | github.com/generalroboticslab ↗ |
| Duke Humanoid v1 | walk (training checkpoints) | Isaac Gym + rl_games | not stated | in source | weights | github.com/generalroboticslab ↗ |
| Fourier N1 | Walk (TorchScript, developer API, radians) | PyTorch JIT (Isaac Gym / legged_gym) | not stated | in source | weights | github.com/FFTAI ↗ |
| Fourier N1 | Walk, refined (TorchScript) | PyTorch JIT (Isaac Gym / legged_gym, FourierN1_refine branch of Wiki-GRx-Gym) | not stated | in source | weights | github.com/FFTAI ↗ |
| Fourier N1 | Built-in tasks: stand/walk/run (up to 3.5 m/s), get-up front/back, squat, single-leg stand, lie down, box-carry arm pose | fourier-grx (closed binary) | not stated | in source | recipe | github.com/FFTAI ↗ |
| HopeJR | hand manipulation via ACT (train-your-own recipe) | LeRobot (ACT) | not stated | in source | recipe | huggingface.co/docs ↗ |
| InMoov | Scripted gestures, speech and chatbot (hand-authored, not learned) | MyRobotLab InMoov2 (Python/Jython scripts + AIML) | not stated | in source | recipe | github.com/MyRobotLab ↗ |
| InMoov | Per-subassembly test scripts (hand, arm, head, torso, finger starter) | MyRobotLab (older InMoov service) | not stated | in source | recipe | github.com/MyRobotLab ↗ |
| K-Bot | Standing (stand_frozen.kinfer) | kinfer | not stated | in source | recipe | github.com/kscalelabs ↗ |
| K-Bot | Walking with keyboard velocity commands (eloquent_ride.kinfer) | kinfer | not stated | in source | recipe | github.com/kscalelabs ↗ |
| K-Bot | Full-body joystick RL control (train + convert to .kinfer) | ksim (JAX/MuJoCo MJX) + kinfer | not stated | in source | recipe | github.com/kscalelabs ↗ |
| K-Bot | K-Sim K-Bot walking training example | ksim (JAX/MuJoCo MJX) | not stated | in source | recipe | github.com/kscalelabs ↗ |
| LEAP Hand | in-hand cube reorientation (LeapHandRot) | Isaac Gym Preview 4 + rl_games PPO (LEAP_Hand_Sim) | not stated | in source | weights | github.com/leap-hand ↗ |
| LeKiwi | LeKiwi pick and place (community SmolVLA fine-tune) | LeRobot (SmolVLA) | not stated | in source | weights | huggingface.co/Sounness ↗ |
| Poppy Humanoid | Stand / sit posture (scripted primitives) | pypot primitives (Python) | not stated | in source | recipe | github.com/poppy-project ↗ |
| Poppy Humanoid | Dance beat, upper-body and head idle motion (scripted) | pypot LoopPrimitive | not stated | in source | recipe | github.com/poppy-project ↗ |
| Poppy Humanoid | Safety: torque limiting and temperature monitoring | pypot LoopPrimitive | not stated | in source | recipe | github.com/poppy-project ↗ |
| Pupper v3 | velocity-tracking locomotion from joystick (default walking policy) | MuJoCo MJX + Brax PPO (pupperv3-mjx), JSON weights run by the neural_controller ROS 2 node | not stated | in source | recipe | github.com/Nate711 ↗ |
| Pupper v3 | three-legged walking | MuJoCo MJX + Brax PPO (pupperv3-mjx) | not stated | in source | recipe | github.com/Nate711 ↗ |
| ROBOTO_ORIGIN | walk (default velocity-tracking policy) | Isaac Lab + RSL-RL (roboparty_train), ONNX via roboparty_inference | not stated | in source | weights | github.com/Roboparty ↗ |
| ROBOTO_ORIGIN | walk, AMP style | Isaac Lab + RSL-RL AMP | not stated | in source | weights | github.com/Roboparty ↗ |
| ROBOTO_ORIGIN | walk, attention encoder | Isaac Lab + RSL-RL | not stated | in source | weights | github.com/Roboparty ↗ |
| ROBOTO_ORIGIN | motion tracking: wave / dance (BeyondMimic) | Isaac Lab BeyondMimic | not stated | in source | weights | github.com/Roboparty ↗ |
| ROBOTO_ORIGIN | get up from the ground | Isaac Lab BeyondMimic-style motion tracking | not stated | in source | weights | github.com/Roboparty ↗ |
| ROBOTO_ORIGIN | interruptible walking | Isaac Lab + RSL-RL | not stated | in source | weights | github.com/Roboparty ↗ |
| ROBOTO_ORIGIN | parkour with depth | Isaac Lab + RSL-RL | not stated | in source | weights | github.com/Roboparty ↗ |
| SO-100 (SO-ARM100) | SmolVLA base (pretrained VLA to fine-tune on your own SO-100 dataset) | LeRobot (SmolVLA) | not stated | in source | weights | huggingface.co/lerobot ↗ |
| SO-100 (SO-ARM100) | SO-100/101 language-conditioned manipulation (VLA, zero-shot or fine-tune) | LeRobot (MolmoAct2) | not stated | in source | weights | huggingface.co/lerobot ↗ |
| SO-101 (SO-ARM101) | Generalist SO-101 pick and place (world action model, fine-tune with the included LoRA recipe) | LeRobot (Flux3Policy) | not stated | in source | weights | huggingface.co/black-forest-la ↗ |
| SO-101 (SO-ARM101) | SO-100/101 language-conditioned manipulation (VLA, zero-shot or fine-tune) | LeRobot (MolmoAct2) | not stated | in source | weights | huggingface.co/lerobot ↗ |
| SO-101 (SO-ARM101) | SmolVLA base (pretrained VLA to fine-tune on your own SO-101 dataset) | LeRobot (SmolVLA) | not stated | in source | weights | huggingface.co/lerobot ↗ |
| ToddlerBot | walk (omnidirectional, 2XC) | MuJoCo MJX + RSL-RL PPO, ONNX Runtime on Jetson | not stated | in source | recipe | drive.google.com/drive ↗ |
| ToddlerBot | get up from the ground | MuJoCo MJX + RSL-RL PPO | not stated | in source | recipe | drive.google.com/drive ↗ |
| ToddlerBot | crawl (full / only / rotate) | MuJoCo MJX + RSL-RL PPO | not stated | in source | recipe | drive.google.com/drive ↗ |
| ToddlerBot | crawl up / down stairs | MuJoCo MJX + RSL-RL PPO | not stated | in source | recipe | drive.google.com/drive ↗ |
| ToddlerBot | climb up / down box, climb wall | MuJoCo MJX + RSL-RL PPO | not stated | in source | recipe | drive.google.com/drive ↗ |
| ToddlerBot | multi-skill whole-body locomotion (Locomotion Beyond Feet) | RL skill policies + depth skill classifier (FoundationStereo) | not stated | in source | recipe | github.com/hshi74 ↗ |
| ToddlerBot | push-up (keyframe replay) | open-loop keyframe replay (run_policy.py --policy replay) | not stated | in source | recipe | github.com/hshi74 ↗ |
| ToddlerBot | cartwheel (DeepMimic-style) | MuJoCo MJX + RSL-RL PPO | not stated | in source | recipe | github.com/hshi74 ↗ |
| ToddlerBot | bimanual manipulation (diffusion policy) | Diffusion Policy (PyTorch) | not stated | in source | recipe | github.com/hshi74 ↗ |
| UMI Gripper (Universal Manipulation Interface) | In-the-wild cup arrangement (rotate espresso cup, place on saucer) | Diffusion Policy (UMI), fine-tuned CLIP ViT-L backbone | not stated | in source | weights | real.stanford.edu/umi ↗ |
| Z-Bot (Zeroth Bot) | Z-Bot joystick/velocity walking (RL benchmark template) | ksim (JAX/MuJoCo) + kinfer | not stated | in source | recipe | github.com/kscalelabs ↗ |
| Z-Bot (Zeroth Bot) | Basic Z-Bot walking policy | ksim + kinfer | not stated | in source | recipe | github.com/kscalelabs ↗ |
| Z-Bot (Zeroth Bot) | Standing policy (Zeroth-01, TPU) | PyTorch -> CVITEK TPU (Milk-V) | not stated | in source | weights | github.com/kscalelabs ↗ |
| Z-Bot (Zeroth Bot) | Demo inference (Zeroth-01 first run) | Python SDK | not stated | in source | recipe | github.com/kscalelabs ↗ |
| Z-Bot (Zeroth Bot) | Scripted demos: wave, salute; voice conversation | kos-zbot | not stated | in source | recipe | github.com/kscalelabs ↗ |