Awesome Physical AI

This repository provides a curated list of academic papers and resources on Physical AI — focusing on Foundation models, Vision-Language-Action (VLA) models, World models, Embodied AI and Resources of Relation Research and Application .

Contributions are welcome! Please feel free to submit pull requests or reach out via email to add papers! If you find this repository useful, please consider citing and giving this list a star ⭐. Feel free to share it with others!

---

Table of Contents

• Foundations
  • Vision-Language Backbones
  • Visual Representations
• VLA Architectures
  • End-to-End VLAs
  • Modular VLAs
  • Compact & Efficient VLAs
• Action Representation & Policy Decoding
  • Action Tokenization & Representation Formats
  • Autoregressive & Discrete Decoding
  • Discrete Diffusion VLAs
  • Continuous Diffusion & Flow Matching Policies
• World Models
  • JEPA & Latent Prediction
  • Generative World Models
  • Embodied World Models
• Reasoning & Planning
  • Chain-of-Thought & Deliberation
  • Error Detection & Recovery
• Learning Paradigms
  • Imitation Learning
  • Reinforcement Learning & Reward Design
• Data Curation & Synthesis
• Scaling & Generalization
  • Scaling Laws
  • Cross-Embodiment Transfer
  • Open-Vocabulary Generalization
• Deployment
  • Quantization & Compression
  • Real-Time Control
• Safety & Alignment
• Lifelong Learning
• Applications
  • Humanoid Robots
  • Manipulation
  • Navigation
• Sim-to-Real Transfer
• Surveys
• Resources
  • Datasets & Benchmarks
  • Simulation Platforms
• Companies & Projects
• Related Works

---

Foundations

Core perception and representation models that serve as the bedrock for Physical AI systems.

Vision-Language Backbones

Foundational VLMs aligning vision and language, widely used as pretrained backbones for robotics.

• CLIP: "Learning Transferable Visual Models From Natural Language Supervision", ICML 2021. [Paper] [Code]
• SigLIP: "Sigmoid Loss for Language Image Pre-Training", ICCV 2023. [Paper]
• PaLI-X: "PaLI-X: On Scaling up a Multilingual Vision and Language Model", CVPR 2024. [Paper]
• LLaVA: "Visual Instruction Tuning", NeurIPS 2023. [Paper] [Project]
• Prismatic VLMs: "Investigating the Design Space of Visually-Conditioned Language Models", ICML 2024. [Paper] [Code]

Visual Representations

Self-supervised encoders and dense perception models for robotic state estimation.

• DINOv2: "Learning Robust Visual Features without Supervision", arXiv, Apr 2023. [Paper] [Code]
• SAM: "Segment Anything", ICCV 2023. [Paper] [Project]
• R3M: "A Universal Visual Representation for Robot Manipulation", CoRL 2022. [Paper] [Code]
• MVP: "Masked Visual Pre-training for Motor Control", arXiv, Mar 2022. [Paper] [Project]
• Grounding DINO: "Marrying DINO with Grounded Pre-Training for Open-Set Object Detection", ECCV 2024. [Paper] [Code]

---

VLA Architectures

End-to-End VLAs

Monolithic models processing vision, language, and actions in a unified architecture (often as interleaved tokens).

• RT-1: "Robotics Transformer for Real-World Control at Scale", RSS 2023. [Paper] [Project]
• RT-2: "Vision-Language-Action Models Transfer Web Knowledge to Robotic Control", CoRL 2023. [Paper] [Project]
• OpenVLA: "An Open-Source Vision-Language-Action Model", CoRL 2024. [Paper] [Project] [Code]
• PaLM-E: "An Embodied Multimodal Language Model", ICML 2023. [Paper] [Project]
• VIMA: "General Robot Manipulation with Multimodal Prompts", ICML 2023. [Paper] [Project]
• Gato: "A Generalist Agent", TMLR 2022. [Paper]
• RoboFlamingo: "Vision-Language Foundation Models as Effective Robot Imitators", ICLR 2024. [Paper]
• Magma: "A Foundation Model for Multimodal AI Agents", arXiv, Feb 2025. [Paper] [Code]
• RoboVLMs: "Towards Generalist Robot Policies", arXiv, Dec 2024. [Paper]
• π0.5 / π0.6: Physical Intelligence, 2025. [Paper] [Blog] (See also: Continuous Policies for flow-matching core)
• GR-3: ByteDance Seed, Jul 2025. [Paper]
• UniVLA: "Unified Vision-Language-Action Model", arXiv, Jun 2025. [Paper] [Code]
• SpatialVLA: "Exploring Spatial Representations for Visual-Language-Action Model", arXiv, Jan 2025. [Paper]
• GraspVLA: "A Grasping Foundation Model Pre-trained on Billion-scale Synthetic Action Data", CoRL 2025. [Paper]
• VLA-0: "Building State-of-the-Art VLAs with Zero Modification", NVIDIA, 2025. [Paper]
• ThinkAct: "Vision-Language-Action Reasoning via Reinforced Visual Latent Planning", NeurIPS 2025. [Paper]
• OmniVLA: "Physically-Grounded Multimodal VLA with Unified Multi-Sensor Perception", Microsoft, 2025. [Paper]
• RoboBrain: "A Unified Brain Model for Robotic Manipulation", CVPR 2025. [Paper]
• DexGraspVLA: "Towards General Dexterous Grasping", PsiBot, 2025. [Paper]
• Motus: "A Unified Latent Action World Model", arXiv, Dec 2024. [Paper]
• InternVLA-A1/M1: InternRobotics, 2025-2026. [Paper] [Paper]
• FLOWER: "Flow-based VLA for CALVIN Benchmark", arXiv, 2025.
• EO-1: "EmbodiedOneVision: Interleaved Vision-Text-Action Pretraining", arXiv, Aug 2025. [Paper]
• LingBot-VLA: "A Pragmatic VLA Foundation Model", Ant Group, Jan 2026. [Paper]

Modular VLAs

Decoupling high-level semantic reasoning (VLM) from low-level motor execution (specialized policies).

• CogACT: "Synergizing Cognition and Action", arXiv, Nov 2024. [Paper]
• Gemini Robotics: "Bringing AI into the Physical World", arXiv, Mar 2025. [Paper] (See also: CoT section for thinking)
• Helix: "A Vision-Language-Action Model for Generalist Humanoid Control", arXiv, Apr 2025.
• SayCan: "Do As I Can, Not As I Say", CoRL 2022. [Paper]
• Code as Policies: "Language Model Programs for Embodied Control", arXiv, Sep 2022. [Paper]
• Inner Monologue: "Embodied Reasoning through Planning with Language Models", CoRL 2022. [Paper]
• HybridVLA: "Collaborative Diffusion and Autoregression", arXiv, Mar 2025. [Paper]
• OpenHelix: "Open-Source Dual-System VLA Model", arXiv, May 2025. [Paper]
• OneTwoVLA: "Unified VLA with Adaptive Reasoning", arXiv, May 2025. [Paper]
• DualVLA: "Partial Decoupling of Reasoning and Action", arXiv, Nov 2025. [Paper]

Compact & Efficient VLAs

Lightweight architectures for edge deployment and fast inference.

• TinyVLA: "Towards Fast, Data-Efficient Vision-Language-Action Models", arXiv, Apr 2025. [Paper]
• SmolVLA: "A Small Vision-Language-Action Model", arXiv, Jun 2025. [Paper] [Code]
• BitVLA: "1-bit Vision-Language-Action Models", arXiv, Jun 2025. [Paper]
• MoLe-VLA: "Dynamic Layer-skipping via Mixture-of-Layers", arXiv, Mar 2025. [Paper]
• NORA / NORA-1.5: "Small Open-Sourced Generalist VLA", arXiv, 2025. [Paper] [Paper]
• DynamicVLA: "Compact 0.4B VLA for Dynamic Object Manipulation", arXiv, Jan 2026. [Paper]

---

Action Representation & Policy Decoding

Note: This section is split to clearly distinguish how actions are formatted (Representation) from how policies generate them (Decoding/Architecture).

Action Tokenization & Representation Formats

Converting continuous motor commands into structured formats (discrete tokens, chunks, or latent codes).

• FAST: "Efficient Action Tokenization via Frequency-Space Compression", arXiv, Jan 2025. [Paper]
• FASTer: "Learnable Action Tokenizer with RVQ and Frequency L1 Loss", ICLR 2026 Submission.
• OmniSAT: "B-Spline-based Compact Action Tokenization", ICLR 2026 Submission.
• ACT: "Action Chunking with Transformers for Bimanual Manipulation", RSS 2023. [Paper] [Code]
• VQ-VLA: "Scaling Vector-Quantized Action Tokenizers", ICCV 2025. [Paper]
• Behavior Transformers: "Cloning k Modes with One Stone", NeurIPS 2022. [Paper]

Autoregressive & Discrete Decoding

VLAs that generate discrete action tokens sequentially (next-token prediction).

• Note: Most End-to-End VLAs (RT-2, OpenVLA, UniVLA) fall into this category by default. Key innovations in this sub-area include:
• GR-1 / GR-2: "Generative Video-Language-Action Models", ICLR 2024 / arXiv 2024. [Paper] [Paper]
• LAPA: "Latent Action Pretraining from Videos", arXiv, Oct 2024. [Paper]
• Actions as Language: "Fine-Tuning VLMs into VLAs Without Catastrophic Forgetting", ICLR 2026 Submission.

Discrete Diffusion VLAs

Using discrete diffusion models to generate action tokens in parallel, speeding up inference over AR.

• Discrete Diffusion VLA: "Bringing Discrete Diffusion to Action Decoding", arXiv, Aug 2025. [Paper]
• dVLA: "Diffusion VLA with Multimodal Chain-of-Thought", ICLR 2026 Submission.
• DIVA: "Discrete Diffusion for Parallelized Action Generation", ICLR 2026 Submission.
• Unified Diffusion VLA: "Joint Discrete Denoising for Frames and Actions", ICLR 2026 Submission.
• LLaDA-VLA: "Vision Language Diffusion Action Models", arXiv, Sep 2025. [Paper]

Continuous Diffusion & Flow Matching Policies

Generating continuous trajectories (rather than discrete tokens) via diffusion or flow matching for high-frequency, dexterous control.

• π₀ (pi-zero): "A Vision-Language-Action Flow Model", arXiv, Oct 2024. [Paper]
• Diffusion Policy: "Visuomotor Policy Learning via Action Diffusion", RSS 2023. [Paper] [Code]
• Octo: "An Open-Source Generalist Robot Policy", RSS 2024. [Paper]
• RDT-1B: "A Diffusion Foundation Model for Bimanual Manipulation", arXiv, Oct 2024. [Paper]
• DexVLA: "VLM with Plug-In Diffusion Expert", arXiv, Feb 2025. [Paper]
• Consistency Policy: "Accelerated Visuomotor Policies via Consistency Distillation", RSS 2024. [Paper]
• Dita: "Scaling Diffusion Transformer for Generalist Policy", arXiv, Mar 2025. [Paper]
• Streaming Flow Policy: "Simplifying diffusion/flow-matching policies", CoRL 2025 Oral. [Paper]
• ManiFlow: "General Robot Policy via Consistency Flow Training", CoRL 2025. [Paper]
• GPC: "Test-time Distribution-level Composition of Flow/Diffusion Policies", arXiv, Oct 2025. [Paper]

---

World Models

Internal models of the environment that enable prediction, planning, and simulation in the mind of the agent. (Moved before Reasoning to support "Simulate then Reason" flow)

JEPA & Latent Prediction

Predicting abstract latent states (not pixels) for efficient world modeling.

• "A Path Towards Autonomous Machine Intelligence", Meta AI, Jun 2022. [Paper]
• I-JEPA / V-JEPA: "Joint-Embedding Predictive Architecture", CVPR 2023 / arXiv 2024. [Paper] [Paper]
• V-JEPA 2: "Self-Supervised Video Models Enable Understanding, Prediction and Planning", Meta, Jun 2025. [Paper]

Generative World Models

Generating pixels, video, or interactive 3D environments.

• World Models: NeurIPS 2018. [Paper]
• DreamerV3: "Mastering Diverse Domains through World Models", arXiv, Jan 2023. [Paper]
• Genie / Genie 2: "Generative Interactive Environments", ICML 2024 / DeepMind 2024. [Paper] [Blog]
• Sora: "Video Generation Models as World Simulators", OpenAI, Feb 2024. [Blog]
• DIAMOND: "Diffusion for World Modeling", NeurIPS 2024. [Paper]
• Cosmos-Predict2.5: "World Simulation with Video Foundation Models", NVIDIA, Oct 2025. [Paper]

Embodied World Models

World models specifically tailored for robotic manipulation and physical reasoning.

• Structured World Models: "Structured World Models from Human Videos", RSS 2023. [Paper]
• WHALE: "Towards Generalizable World Models for Embodied Decision-making", arXiv, Nov 2024. [Paper]
• UniSim: "Learning Interactive Real-World Simulators", ICLR 2024. [Paper]
• PointWorld: "Scaling 3D World Models for In-The-Wild Manipulation", arXiv, Jan 2026. [Paper]
• DreamDojo: "Generalist Robot World Model from Large-Scale Human Videos", arXiv, Feb 2026. [Paper]

---

Reasoning & Planning

Chain-of-Thought & Deliberation

Explicit "thinking" processes before acting, including value-guided search and spatial reasoning.

• Hume: "Introducing System-2 Thinking in Visual-Language-Action Model", arXiv, May 2025. [Paper]
• Embodied-CoT: "Robotic Control via Embodied Chain-of-Thought Reasoning", arXiv, Jul 2024. [Paper]
• ReAct: "Synergizing Reasoning and Acting in Language Models", ICLR 2023. [Paper]
• ReKep: "Spatio-Temporal Reasoning of Relational Keypoint Constraints", CoRL 2024. [Paper]
• TraceVLA: "Visual Trace Prompting Enhances Spatial-Temporal Awareness", arXiv, Dec 2024. [Paper]
• Cosmos-Reason1: "From Physical Common Sense To Embodied Reasoning", NVIDIA, Mar 2025. [Paper]
• IPR-1: "Interactive Physical Reasoner", CVPR 2026. [Paper]
• MEM: "Multi-Scale Embodied Memory for Long-Horizon Tasks", Physical Intel, Mar 2026. [Paper]
• InstructVLA: "Instruction Tuning: From Understanding to Manipulation", arXiv, Jul 2025. [Paper]

Error Detection & Recovery

Real-time failure detection and correction.

• DoReMi: "Detecting and Recovering from Plan-Execution Misalignment", arXiv, Jul 2023. [Paper]
• Code-as-Monitor: "Constraint-aware Visual Programming for Failure Detection", CVPR 2025. [Paper]
• AHA: "Detecting and Reasoning over Failures", arXiv, Oct 2024.

---

Learning Paradigms

Imitation Learning

Behavioral cloning and learning from demonstrations (human or robot).

• CLIPort: "What and Where Pathways for Manipulation", CoRL 2021. [Paper]
• MimicPlay: "Long-Horizon Imitation Learning by Watching Human Play", CoRL 2023. [Paper]
• RVT / RVT-2: "Robotic View Transformer", CoRL 2023 / RSS 2024. [Paper] [Paper]
• Perceiver-Actor: "Multi-Task Transformer for Manipulation", CoRL 2022. [Paper]
• Phantom: "Training Robots Without Robots Using Only Human Videos", arXiv, Mar 2025. [Paper]
• ZeroMimic: "Distilling Skills from Web Videos", arXiv, Mar 2025. [Paper]
• DataMIL: "Selecting Data for Imitation Learning", arXiv, May 2025. [Paper]

Reinforcement Learning & Reward Design

Optimizing policies via trial-and-error, including automated reward generation from language models. (Reward Design merged here)

• CO-RFT: "Chunked Offline Reinforcement Learning Fine-Tuning", arXiv, 2026.
• FLaRe: "Achieving Masterful Policies with Large-Scale RL Fine-Tuning", arXiv, Sep 2024. [Paper]
• RL4VLA: "What Can RL Bring to VLA Generalization?", NeurIPS 2025.
• VLA-RL: "Towards Masterful Manipulation with Scalable RL", arXiv, May 2025. [Paper]
• RoboMonkey: "Scaling Test-Time Sampling and Verification", CoRL 2025. [Paper]
• Embodied-R1: "Reinforced Embodied Reasoning", arXiv, Aug 2025. [Paper]
• Residual RL for VLAs: "Self-Improvement via Residual RL", ICLR 2026 Submission.
• Text2Reward: "Automated Dense Reward Function Generation", arXiv, Sep 2023. [Paper] (Reward Design)
• Language to Rewards: "Robotic Skill Synthesis via Rewards", CoRL 2023. [Paper] (Reward Design)

---

Data Curation & Synthesis

(New Section) Generating, augmenting, and curating robotic datasets using foundation models to enable scaling.

• GenSim / GenSim2: "Procedural Task Generation for Robot Learning", arXiv, 2023. [Paper]
• RoboGen: "A Generative and Self-Guided Robotic Agent", arXiv, Nov 2023. [Paper]
• AutoRT: "Embodied Foundation Models for Large Scale Orchestration", ICRA 2024. [Paper]
• RE3SIM: "Generating High-Fidelity Simulation Data via 3D-Photorealistic Real-to-Sim", arXiv, Feb 2025. [Paper]

---

Scaling & Generalization

Scaling Laws

Mathematical relationships governing model/data scale and performance.

• "Neural Scaling Laws for Embodied AI", arXiv, May 2024. [Paper]
• "Data Scaling Laws in Imitation Learning", arXiv, Oct 2024. [Paper]
• EgoScale: "Scaling Dexterous Manipulation with Egocentric Human Data", arXiv, Feb 2026. [Paper]

Cross-Embodiment Transfer

Single policies generalizing across diverse robot morphologies.

• RT-X: "Open X-Embodiment Dataset and RT-X Models", ICRA 2024. [Paper]
• GENBOT-1K: "Training on ~1000 Robot Bodies", arXiv, 2025.
• Crossformer: "One Policy for Manipulation, Navigation, Locomotion", CoRL 2024. [Paper]
• HPT: "Heterogeneous Pre-trained Transformers", NeurIPS 2024. [Paper]
• X-VLA: "Soft-Prompted Transformer for Cross-Embodiment", arXiv, Oct 2025. [Paper]
• HiMoE-VLA: "Hierarchical Mixture-of-Experts for Embodiment Adaptation", arXiv, Dec 2025. [Paper]
• Being-H0.5: "Human-Centric VLA for Cross-Embodiment", arXiv, Jan 2026. [Paper]

Open-Vocabulary Generalization

Handling novel visual appearances, objects, and semantic concepts unseen during training.

• MOO: "Open-World Object Manipulation", CoRL 2023. [Paper]
• VoxPoser: "Composable 3D Value Maps from Language", CoRL 2023. [Paper]
• RoboPoint: "Spatial Affordance Prediction", CoRL 2024. [Paper]
• CLIP-Fields: "Weakly Supervised Semantic Fields", RSS 2023. [Paper]
• VLMaps: "Visual Language Maps for Navigation", ICRA 2023. [Paper]

---

Deployment

Quantization & Compression

Low-bit quantization and model pruning for edge robots.

• BitVLA: "1-bit Vision-Language-Action Models", arXiv, 2025.
• DeeR-VLA: "Dynamic Inference for Efficient Robot Execution", arXiv, Nov 2024. [Paper]
• AutoQVLA: "Channel-Aware Quantization for OpenVLA", ICLR 2026 Submission.
• HyperVLA: "Hypernetworks for Efficient Inference", ICLR 2026 Submission.
• SpecPrune-VLA: "Action-Aware Self-Speculative Pruning", arXiv, Sep 2025. [Paper]

Real-Time Control

Bridging high-latency AI inference with low-latency physical control loops.

• Real-Time Chunking: "Asynchronous Execution for Robot Control", Physical Intel, 2025. [Paper]
• A2C2: "Asynchronous Action Chunk Correction", arXiv, 2025.

---

Safety & Alignment

Guardrails, ethics, and robustness against adversarial inputs.

• Robot Constitution: "Data-driven Natural Language Rules", Gemini Robotics, 2025. [Paper]
• RoboPAIR: "Jailbreaking LLM-Controlled Robots", ICRA 2025. [Paper]
• RoboGuard: "Safety Guardrails for LLM-Enabled Robots", arXiv, Apr 2025.
• "Robots Enact Malignant Stereotypes", FAccT 2022. [Paper]
• "Highlighting Safety Concerns of Deploying VLMs in Robotics", arXiv, Feb 2024. [Paper]

---

Lifelong Learning

Continuous adaptation and skill accumulation without catastrophic forgetting.

• Voyager: "An Open-Ended Embodied Agent with LLMs", arXiv, May 2023. [Paper]
• RoboCat: "A Self-Improving Generalist Agent", arXiv, Jun 2023. [Paper]
• JARVIS-1: "Open-World Multi-task Agents with Memory-Augmented LLMs", arXiv, Nov 2023. [Paper]
• LOTUS: "Continual Imitation Learning via Unsupervised Skill Discovery", arXiv, Dec 2024.

---

Applications

Humanoid Robots

• GR00T N1: "An Open Foundation Model for Humanoid Robots", NVIDIA, Mar 2025. [Paper]
• HumanPlus: "Humanoid Shadowing and Imitation from Humans", arXiv, Jun 2024. [Paper]
• OmniH2O: "Universal Human-to-Humanoid Teleoperation", CoRL 2024. [Paper]
• WholeBodyVLA: "Unified Latent VLA for Loco-Manipulation", ICLR 2026. [Paper]

Manipulation

• Manipulate-Anything: "Automating Real-World Robots using VLMs", CoRL 2024. [Paper]
• A3VLM: "Actionable Articulation-Aware Vision Language Model", CoRL 2024. [Paper]
• LaN-Grasp: "Language-Driven Grasp Detection", CVPR 2024. [Paper]

Navigation

• LM-Nav: "Robotic Navigation with Large Pre-Trained Models", CoRL 2022. [Paper]
• NaVILA: "Legged Robot VLA for Navigation", arXiv, Dec 2024. [Paper]
• NaVid: "Video-based VLM for Vision-and-Language Navigation", RSS 2024. [Paper]
• CoW: "CLIP on Wheels: Zero-Shot Object Navigation", ICRA 2023. [Paper]
• WMNav: "Integrating VLMs into World Models for Navigation", arXiv, Mar 2025. [Paper]
• InternNav: "Dual-System Foundation Model for VLN", arXiv, Dec 2025. [Paper]
• OmniVLA-Nav: "Omni-Modal VLA for Robot Navigation", arXiv, Sep 2025. [Paper]

---

Sim-to-Real Transfer

Bridging the gap between simulation training and real-world deployment.

• RE3SIM: "High-Fidelity Real-to-Sim Data Generation", arXiv, Feb 2025. [Paper]
• Real2Render2Real: "Scaling Robot Data Without Dynamics Simulation", arXiv, May 2025. [Paper]
• RSR Loop: "Real-Sim-Real Loop with Differentiable Simulation", arXiv, Mar 2025. [Paper]
• "Sim-to-Real Reinforcement Learning for Vision-Based Dexterous Manipulation", arXiv, Feb 2025. [Paper]

---

Surveys

• Foundation Models in Robotics: IJRR 2025. [Paper]
• VLA Anatomy Survey: "From Modules to Milestones", arXiv, Dec 2025. [Paper]
• Pure VLA Survey: "Comprehensive Taxonomy", arXiv, Sep 2025. [Paper]
• World Models for Embodied AI: arXiv, Oct 2025. [Paper]
• Diffusion Policy Survey: TechRxiv 2025. [Paper]
• Action Tokenization Survey: "VLA Survey: An Action Tokenization Perspective", arXiv, Jul 2025. [Paper]
• VLA Post-Training: "Post-Training and Human Motor Learning", arXiv, Jun 2025. [Paper]
• Embodied AI Safety: arXiv 2026. [Paper]

---

Resources

Datasets & Benchmarks

Dataset	Scale	Focus
Open X-Embodiment	1M+ trajectories	Cross-embodiment
DROID	76K trajectories	In-the-wild manipulation
BridgeData V2	Multi-task	Few-shot transfer
LIBERO	130 tasks	Lifelong learning
SIMPLER	Sim-to-real	Policy evaluation
CALVIN	Long-horizon	Language-conditioned
RLBench	100 tasks	Manipulation benchmark
RoboCasa	Large-scale	Household tasks
VLABench	Long-horizon	Reasoning benchmark
RoboMIND	Multi-embodiment	Intelligence benchmark
EWMBench	World model eval	Scene/motion/semantic
ManipBench	VLM evaluation	Low-level manipulation

Simulation Platforms

Platform	Focus
ManiSkill3	GPU-parallelized robotics
Genesis	Differentiable physics
Isaac Lab / Isaac Sim	NVIDIA robotics simulation
OmniGibson	High-fidelity home simulation
Habitat 2.0	Navigation & rearrangement
BEHAVIOR-1K	1,000 everyday activities

---

Companies & Projects

Company	Focus
Physical Intelligence (π)	General-purpose robot foundation models
Google DeepMind	RT, Gemini Robotics, Genie
OpenAI	CLIP, GPT-4V, Sora
Meta AI (FAIR)	JEPA, DINOv2, SAM
World Labs	Spatial intelligence & world models
NVIDIA	GR00T, Isaac Sim, Cosmos
Hugging Face	LeRobot, SmolVLA
Covariant	Industrial robotics AI
Skild AI	General-purpose robot brain
Figure AI / 1X / Tesla	Humanoid robots

---

Related Works

• Awesome World Models
• Awesome-VLA-Robotics
• Awesome-Robotics-Foundation-Models
• Awesome-LLM-Robotics

---

Citation

@misc{awesome-physical-ai,
    title = {Awesome Physical AI},
    author = {Evan},
    journal = {GitHub repository},
    url = {https://github.com/EtachGu/Awesome-Physical-AI},
    year = {2026},
}