humanoidrobots.systems

humanoidrobots.systems/brainsbrains.humanoidrobots.systems

Robot brains: the models behind the bodies.

Vision-language-action models, world models and learned controllers that run humanoids. The openness column states what you can actually get: weights you can download, code only, an API, or nothing. Newest first.

Models
40
Open weights
18
API only
2
Closed
16
Robot foundation models with kind, openness, license, release and embodiments.
ModelCompanyKindAccessLicenseReleasedEmbodimentsLinks
Skild S1Skild AIVLAClosed—18 Aug 2026—site
Gemini Robotics 2Google DeepMindVLAClosed—30 Jul 2026Apptronik Apollo, Franka Duo, The SO-101 (SO-ARM101), Trossen, Dexmatesite
Gemini Robotics ER 2Google DeepMindVLMAPI only—30 Jul 2026—site
Gemini Robotics On-Device 2Google DeepMindVLAClosed—30 Jul 2026—site
Cosmos 3NVIDIAWorld modelOpen weightsLicenseRef-OpenMDW-1.1May 2026—codeweightssite
Isaac GR00T N1.7NVIDIAVLAOpen weightsLicenseRef-NVIDIA-Open-Model17 Apr 2026Unitree G1, AgiBot Genie-1, bimanual YAMpapercodeweights
π0.7Physical IntelligenceVLAClosed—16 Apr 2026—papersite
Genie Operator-2 (GO-2)AgiBot (Zhiyuan Robotics)VLAClosed—9 Apr 2026—site
Helix 02Figure AIVLAClosed—27 Jan 2026Figure 03site
LingBot-VLARobbyant (Ant Group)VLAOpen weightsApache-2.026 Jan 2026dual-arm robotspapercodeweights
1X World Model (1XWM)1X TechnologiesWorld modelClosed—12 Jan 20261X NEOsite
Isaac GR00T N1.6NVIDIAVLAOpen weightsLicenseRef-NVIDIA-OneWay-NoncommercialDec 2025Unitree G1, AgiBot Genie-1, bimanual YAM, Fourier GR-1codeweightssite
π0.6 / π*0.6Physical IntelligenceVLAClosed—17 Nov 2025—papersite
Gemini Robotics 1.5Google DeepMindVLAClosed—25 Sep 2025Stanford ALOHA, Franka bi-arm, Apptronik Apollopapersite
Gemini Robotics-ER 1.5Google DeepMindVLMAPI only—25 Sep 2025—papersite
Skild BrainSkild AIVLAClosed—29 Jul 2025quadrupeds, humanoids, table-top arms, mobile manipulatorssite
Large Behavior Model (LBM 1.0)Toyota Research InstituteVLAClosed—7 Jul 2025Franka FR3 bimanual, Boston Atlas (electric)papersite
Gemini Robotics On-DeviceGoogle DeepMindVLAClosed—24 Jun 2025Stanford ALOHA, Franka FR3 bi-arm, Apptronik Apollocodesite
V-JEPA 2Meta FAIRWorld modelOpen weightsMIT11 Jun 2025Frankapapercodeweights
Redwood AI1X TechnologiesVLAClosed—10 Jun 20251X NEOsite
SmolVLAHugging FaceVLAOpen weightsApache-2.02 Jun 2025The SO-100 (SO-ARM100), The SO-101 (SO-ARM101)papercodeweightssite
Isaac GR00T N1.5NVIDIAVLAOpen weightsLicenseRef-NVIDIA-OneWay-NoncommercialJun 2025Fourier GR-1, Unitree G1, The SO-100 (SO-ARM100)papercodeweightssite
GR00T-Dreams (DreamGen)NVIDIAOtherOpen codeApache-2.019 May 2025Fourier GR-1, Franka, The SO-100 (SO-ARM100)papercodesite
UniVLAOpenDriveLab (HKU)VLAOpen weightsApache-2.09 May 2025—papercodeweights
π0.5 (pi05)Physical IntelligenceVLAOpen weightsApache-2.022 Apr 2025mobile manipulators, Franka (DROID)papercode
Isaac GR00T N1NVIDIAVLAOpen weightsLicenseRef-NVIDIA-OneWay-Noncommercial18 Mar 2025Fourier GR-1, Franka, bimanual armspapercodeweights
Gemini RoboticsGoogle DeepMindVLAClosed—12 Mar 2025Stanford ALOHA, Franka bi-arm, Apptronik Apollopapersite
Genie Operator-1 (GO-1)AgiBot (Zhiyuan Robotics)VLAOpen weightsCC-BY-NC-SA-4.09 Mar 2025AgiBot Genie-1papercodeweightssite
HelixFigure AIVLAClosed—20 Feb 2025Figure 02site
MagmaMicrosoft ResearchVLAOpen weightsMIT18 Feb 2025WidowXpapercodeweightssite
π0-FASTPhysical IntelligenceVLAOpen weightsApache-2.016 Jan 2025Franka (DROID), Stanford ALOHApapercode
Cosmos World Foundation Models (Predict 1 to 2.5)NVIDIAWorld modelOpen weightsLicenseRef-NVIDIA-Open-Model7 Jan 2025—papercodeweights
π0 (pi0)Physical IntelligenceVLAOpen weightsApache-2.031 Oct 2024Stanford ALOHA, Franka (DROID), UR5e, bimanual Trossen, mobile manipulatorspapercode
RDT-1B (Robotics Diffusion Transformer)Tsinghua University (TSAIL)VLAOpen weightsMIT10 Oct 2024AgileX bimanual (ALOHA-style)papercodeweightssite
OpenVLAStanford UniversityVLAOpen weightsMIT13 Jun 2024WidowX, Frankapapercodeweightssite
OctoUC Berkeley Hybrid Robotics LabVLAOpen weightsMIT20 May 2024WidowX, Franka, UR5papercodeweightssite
RT-2Google DeepMindVLAClosed—28 Jul 2023Everyday Robots mobile manipulatorpapersite
ACT (Action Chunking with Transformers)Stanford UniversityVLAOpen codeMIT23 Apr 2023Stanford ALOHApapercodesite
Diffusion PolicyToyota Research InstituteVLAOpen codeMIT7 Mar 2023UR5, Frankapapercodesite
RT-1 (Robotics Transformer)Google DeepMindVLAOpen codeApache-2.013 Dec 2022Everyday Robots mobile manipulatorpapercodesite