Compare
VLA models
Compare robotic foundation models (VLA): parameters, architecture, open weights, control frequency. Sourced data.
To understand how these models work before comparing them, read our VLA models explained guide.
Click any row to reveal details, sources and verification date.
| Name | Organization | Country | Release | Params (B) | Open weights | Architecture | Control (Hz) |
|---|---|---|---|---|---|---|---|
| NVIDIA Isaac GR00T N1.7 | NVIDIA | US | 2026-04 | 3 | yes | dual-system VLA: Cosmos-Reason2-2B VLM (System 2) + 16-layer diffusion-transformer action head (System 1), 40-step action horizon, relative end-effector action space | n/a |
|
NVIDIA's flagship open robot foundation model line, started with GR00T N1 (first open humanoid foundation model, March 2025). N1.7 (GA April 2026) swaps in a Cosmos-Reason2-2B backbone and a relative end-effector action space shared between robot and human embodiments.
License : NVIDIA Open Model License (weights); Apache-2.0 (code)Modalities : vision, language, actionEmbodiments : humanoid, manipulator, cross-embodimentTraining data : Several thousand hours of teleoperated demos (bimanual YAM, AgiBot Genie-1, simulated Galaxea R1 Pro, Unitree G1 locomanipulation) plus 20,000 h of EgoScale egocentric human video
Sources : NVIDIA (GitHub) (2026-07-09)NVIDIA Research (GEAR Lab) (2025-12-15)arXiv (2025-03-18) Last verified 2026-07-09 |
|||||||
| Gemini Robotics 1.5 (+ Gemini Robotics-ER 1.5) | Google DeepMind | US | 2025-09 | n/a | no | dual-model agentic system: Gemini Robotics-ER 1.5 (embodied-reasoning orchestrator with tool calling) + Gemini Robotics 1.5 VLA executor with motion transfer and interleaved natural-language thinking | n/a |
|
Google DeepMind's agentic robotics stack: ER 1.5 plans, reasons spatially and calls tools (available via the Gemini API; ER 1.6 followed in 2026), while the Gemini Robotics 1.5 VLA executes and transfers motions across embodiments. Model weights remain closed, with the VLA limited to selected partners.
Modalities : vision, language, actionEmbodiments : manipulator, humanoid, cross-embodimentTraining data : Not disclosed in detail; multi-embodiment robot data across ALOHA 2, bi-arm Franka and Apptronik Apollo, leveraged via a Motion Transfer mechanism on top of Gemini multimodal pretraining
Sources : Google DeepMind (2025-09-25)arXiv (2025-10-03)Google AI for Developers (2026-07-09) Last verified 2026-07-09 |
|||||||
| π*0.6 (Physical Intelligence) | Physical Intelligence | US | 2025-11 | n/a | no | VLA built on a 5B-parameter VLM augmented with an action expert; RL post-training via RECAP (RL with Experience and Corrections via Advantage-conditioned Policies) | n/a |
|
Latest VLA from the best-funded US robot foundation model startup, trained with RECAP to learn from its own experience; ran espresso making, laundry folding and factory box packing for hours uninterrupted. Unlike π0/π0.5 (open-sourced via openpi), π*0.6 weights had not been released as of mid-2026.
Modalities : vision, language, actionEmbodiments : manipulator, mobile, cross-embodimentTraining data : Builds on the π0.5 pretraining mixture; RECAP combines demonstrations, expert teleoperated corrections and autonomous on-robot experience with a value function
Sources : Physical Intelligence (2025-11-17)Physical Intelligence (model card) (2025-11-17)Physical Intelligence (GitHub) (2026-07-09) Last verified 2026-07-09 |
|||||||
| Helix | Figure AI | US | 2025-02 | 7 | no | dual-system VLA (System 1/System 2): 7B internet-pretrained VLM at 7-9 Hz + 80M reactive visuomotor policy at 200 Hz | 200 |
|
First VLA to output continuous high-rate control of the entire humanoid upper body (wrists, individual fingers, torso, head) and to run fully onboard embedded low-power GPUs, including zero-shot two-robot collaboration with a single set of weights.
Modalities : vision, language, actionEmbodiments : humanoidTraining data : ~500 h of teleoperated humanoid manipulation demonstrations with auto-generated natural-language annotations
Sources : Figure AI (2025-02-20)Humanoids Daily (2026-07-09) Last verified 2026-07-09 |
|||||||
| OpenVLA | Stanford / UC Berkeley / TRI / Google DeepMind (academic consortium) | US | 2024-06 | 7 | yes | autoregressive VLA: Prismatic VLM (Llama-2-7B language backbone + fused SigLIP and DINOv2 vision encoders), actions as discrete tokens | n/a |
|
The reference open-source VLA: beat the closed 55B RT-2-X with a 7B model and released all checkpoints and training code under MIT, catalyzing the academic and startup VLA ecosystem.
License : MITModalities : vision, language, actionEmbodiments : manipulator, cross-embodimentTraining data : 970k robot manipulation episodes from the Open X-Embodiment dataset; trained on 64 A100 GPUs for 15 days
Sources : arXiv (2024-06-13)Hugging Face (model card) (2026-07-09)OpenVLA project (2026-07-09) Last verified 2026-07-09 |
|||||||
| Octo | UC Berkeley / Stanford / CMU / Google DeepMind (academic consortium) | US | 2024-05 | 0.1 | yes | transformer policy with diffusion action head; flexible observation and action spaces (Octo-Small 27M, Octo-Base 93M) | n/a |
|
First fully open-source generalist robot policy (data, weights and training pipeline); its compact diffusion-head transformer became a standard academic baseline for cross-embodiment manipulation.
License : MITModalities : vision, language, actionEmbodiments : manipulator, cross-embodimentTraining data : 800k robot trajectories from the Open X-Embodiment dataset
Sources : arXiv (2024-05-20)Octo project (2026-07-09)GitHub (2026-07-09) Last verified 2026-07-09 |
|||||||
| RT-2 (Robotic Transformer 2) | Google DeepMind | US | 2023-07 | 55 | no | VLA co-fine-tuned from web-scale VLMs (PaLI-X 55B and PaLM-E 12B variants); robot actions emitted as text tokens | n/a |
|
The founding VLA: first to show that web-scale vision-language knowledge transfers directly to robot control by treating actions as text tokens. It spawned the RT-X / Open X-Embodiment cross-lab effort that underpins most later open VLAs. Research artifact, never released as weights or product.
Modalities : vision, language, actionEmbodiments : manipulator, mobileTraining data : Internet-scale vision-language data co-fine-tuned with RT-1 robot demonstrations (collected with 13 robots over 17 months)
Sources : Google DeepMind (2023-07-28)arXiv (2023-07-28)InfoQ (2023-10-01) Last verified 2026-07-09 |
|||||||
| SmolVLA | Hugging Face (LeRobot team) | US | 2025-06 | 0.5 | yes | compact VLA: SmolVLM-2 backbone + flow-matching action expert conditioned on multi-camera views, robot state and language; asynchronous inference | n/a |
|
A 450M open VLA trained only on crowd-sourced community data, small enough to train on a single consumer GPU and run on CPU or a MacBook; asynchronous inference yields ~30% faster response and ~2x task throughput. Built by Hugging Face's Paris-based LeRobot team.
License : Apache-2.0Modalities : vision, language, actionEmbodiments : manipulatorTraining data : Exclusively publicly available, crowd-sourced LeRobot community datasets from the Hugging Face hub (low-cost robot data)
Sources : arXiv (2025-06-02)Hugging Face (2025-06-03)Hugging Face (model card) (2026-07-09) Last verified 2026-07-09 |
|||||||
| Robostral Navigate | Mistral AI | FR | 2026-07 | 8 | no | pointing-based navigation model: grounding-specialized Mistral VLM predicts target image coordinates and desired orientation in the current camera view, with local displacement fallback | n/a |
|
Mistral AI's entry into physical AI (announced 8 July 2026): an 8B model steering wheeled, legged and flying robots from a single RGB camera and plain-language instructions, trained entirely in simulation. Availability and licensing had not been announced at launch.
Modalities : vision, language, actionEmbodiments : mobile, legged, aerial, cross-embodimentTraining data : ~400,000 simulated trajectories across 6,000 scenes, generated entirely in-house without open-source VLMs; further improved with online reinforcement learning (+3.2 pts reported)
Sources : Mistral AI (2026-07-08)The Decoder (2026-07-08)Bloomberg (2026-07-08) Last verified 2026-07-09 |
|||||||
| GAIA-2 | Wayve | GB | 2025-03 | n/a | no | latent-diffusion multi-camera generative world model conditioned on ego-vehicle dynamics, agent configurations, environmental factors and road semantics | n/a |
|
Europe's flagship driving world model: generates controllable, spatiotemporally consistent multi-camera video, including rare and safety-critical scenarios for training and validating Wayve's driving AI. Succeeded in December 2025 by GAIA-3 (15B parameters, double GAIA-2's size), offered for AV validation.
Modalities : vision, actionEmbodiments : autonomous-vehicleTraining data : Real-world driving data from geographically diverse environments (UK, US, Germany)
Sources : Wayve (2025-03-26)arXiv (2025-03-26)ADAS & Autonomous Vehicle International (2025-12-03) Last verified 2026-07-09 |
|||||||
| RFM-1 | Covariant | US | 2024-03 | 8 | no | 8B any-to-any multimodal transformer performing autoregressive next-token prediction across text, images, video, robot actions and numerical sensor readings | n/a |
|
Early commercial robotics foundation model for warehouse manipulation, including a learned physics world model that predicts video outcomes of candidate actions. In August 2024 Amazon hired Covariant's founders and about a quarter of its staff and took a non-exclusive license to its models.
Modalities : vision, language, action, video, sensorEmbodiments : manipulatorTraining data : Internet data plus Covariant's proprietary multimodal warehouse production data (Covariant Brain fleet: picking, sortation, induction, depalletization since 2017)
Sources : Covariant (2024-03-11)IEEE Spectrum (2024-03-11)TechCrunch (2024-08-31) Last verified 2026-07-09 |
|||||||
| AgiBot GO-1 (Genie Operator-1) | AgiBot (Zhiyuan Robotics) / OpenDriveLab | CN | 2025-03 | 3 | yes | ViLLA (Vision-Language-Latent-Action): InternVL2.5-2B VLM backbone + MoE with a 24-layer latent planner predicting latent action tokens and a high-frequency action expert | n/a |
|
China's flagship open embodied foundation model, introducing the ViLLA framework where latent action tokens bridge vision-language inputs and robot control. Announced March 2025 and fully open-sourced (weights on Hugging Face) in September 2025, alongside the largest open real-robot dataset.
License : CC BY-NC-SA 4.0Modalities : vision, language, actionEmbodiments : humanoid, manipulator, cross-embodimentTraining data : AgiBot World dataset: over 1 million real-robot trajectories across 217 tasks in 5 application domains, plus cross-embodiment and human video data for the latent planner
Sources : Hugging Face (model card) (2026-07-09)GlobeNewswire (AgiBot press release) (2025-03-11)arXiv (2025-03-09)OpenDriveLab (X) (2025-09-20) Last verified 2026-07-09 |
|||||||
12 / 12 entries