VLA & VLM Robot Model Directory
Side-by-side comparison of 50+ open-source vision-language-action (VLA) and vision-language (VLM) models for robotics — OpenVLA, Octo, RT-X, Diffusion Policy, LeRobot, and more — with benchmarks, licenses, hardware requirements, and paper links.
Foundation VLA Models
Large-scale action models trained with multi-robot, multi-task data.
CollectionModel Comparison Track
Models commonly used for benchmark-style side-by-side evaluation.
CollectionPolicy Learning Models
Architectures optimized for practical IL/RL policy training loops.
High-Intent Model Guides
VLA models for robotics
Foundation action models, trade-offs, and fit.
Workflow GuideTeleop bootstrapping models
What works best when demonstrations are your starting point.
Manipulation GuideContact-rich models
Force, tactile signals, and recovery-aware policy choices.
Decision GuideFoundation vs task policies
Broad capability versus faster narrow deployment.
Decision GuideHow to choose a robot model
Data, task scope, evaluation, and deployment constraints.
OpenArm GuideOpenArm models
Policy choices and practical starting paths for OpenArm.
Our team has deployed OpenVLA, Octo, and RT-X on OpenArm, Unitree G1, and Mobile ALOHA rigs. Tell us your robot and task and we will recommend a model and a dataset.
Get a Model RecommendationPopular Categories
Popular Tags
VLA & VLM Models for Robotics
OpenVLA
7B-parameter VLA. Llama 2 + DINOv2/SigLIP. 970K demos from Open X-Embodiment. Outperforms RT-2-X with 7× fewer params. MIT, Hugging Face.
View model →
Octo
Transformer diffusion policy. 27M/93M params. 800K trajectories. Multi-robot, language/goal conditioning. MIT, Hugging Face.
View model →
RT-X / RT-1-X
Open X-Embodiment models. JAX & TensorFlow checkpoints. Multi-robot, language-conditioned. Apache 2.0.
View model →InternVLA-M1
Spatially guided VLA. Two-stage: grounding + action. 71–81% on Google Robot, 95.9% LIBERO. MIT, Hugging Face.
View model →
RoboFlamingo
OpenFlamingo-based VLM for robot control. Policy head + imitation learning. Strong on CALVIN. MIT, Hugging Face.
View model →
BridgeVLA
3D VLA with input-output alignment. 88.2% RLBench, 64% COLOSSEUM. Heatmap pre-training + point cloud fine-tuning.
View model →
Diffusion Policy
Visuomotor policy as denoising diffusion. +46.9% over prior methods. Receding horizon, time-series transformer. Open source.
View model →
LeRobot
Framework + ACT, SmolVLA (450M). End-to-end IL/RL. Datasets, training, deployment. PyTorch, Hugging Face Hub.
View model →Datasets & Tools to Pair
Practical Model Selection
Compare architectures by task fit, data need, and deployment complexity.
Data-Model Alignment
Model choices are connected to compatible dataset and format stacks.
Experiment Velocity
Open-source links and implementation-ready pointers reduce setup friction.
Scale to Production
From evaluation to deployment with support for tuning and integration.