Identity
- Status
- active
- Developer ID
openvla/openvla-7b- Released
- Jun 13, 2024
- Version
- 7B
- License
- Unknown
OpenVLA Research TeamOpenVLA 7B
The OpenVLA team's open 7B vision-language-action model trained across Open X-Embodiment robot trajectories.
- Model type
- Vision-language-action model
- Inputs
- Text, Image, Robot state
- Inference
- Flexible
- Released
- Jun 13, 2024
Providers & Pricing
0 current hosted offerings
No Hosted API Available
This record is downloadable or self-hosted. Its cost depends on the infrastructure and serving configuration, not a universal model price.
At a Glance
Model Facts
Model Design
- Model type
- Vision-language-action model
- Inputs
- Text, Image, Robot state
- Outputs
- Robot action
- Total parameters
- 7B
- Architecture
- Fused SigLIP and DINOv2 visual encoder with Llama 2 7B backbone
Deployment
- Inference
- Flexible
- Control rate
- Unknown
- API available
- No
- Open weights
- Yes
- Self-hostable
- Yes
Publisher Media
OpenVLA 7B Demonstrations
Embodied Intelligence
Robotics Control Profile
Control Contract
Model typeVision-language-action model
Action representationTokenized actions decoded to continuous robot controls
Control architectureFused SigLIP and DINOv2 visual encoder with Llama 2 7B backbone
InferenceFlexible
Native control rateUnknown
Training Data and Embodiments
970,000 robot manipulation trajectories from Open X-Embodiment described by the authors.
Verified Platform Links
Compatible Robots
No Directly Verified Platform Link YetReported embodiment names remain visible above, but are not promoted to canonical robot relationships without direct evidence.
This page represents one developer model product. Dated API snapshots, serving endpoints, pricing, regions, quantizations, and service tiers attach as versioned aliases or provider details and never create duplicate public model records.
Version Lineage
PredecessorUnknown
SuccessorsUnknown
Aliasesopenvla/openvla-7b
Explicit Unknowns
- Native control rate
- License
Primary Evidence
Sources and Observation Date
Comparable Peers
Related Model Comparisons
| A | Pair | B | Context |
|---|---|---|---|
OpenVLA 7BOpenVLA Research Team | vs |  MolmoAct 2Ai2 | cross-developer peersrobotics |
OpenVLA 7BOpenVLA Research Team | vs |  Helix 02Figure | cross-developer peersrobotics |
OpenVLA 7BOpenVLA Research Team | vs |  Gemini Robotics 2Google DeepMind | cross-developer peersrobotics |
OpenVLA 7BOpenVLA Research Team | vs |  SmolVLA 450MHugging Face LeRobot | cross-developer peersrobotics |
OpenVLA 7BOpenVLA Research Team | vs |  GR00T N1.7 3BNVIDIA | cross-developer peersrobotics |
OpenVLA 7BOpenVLA Research Team | vs |  pi 0.7Physical Intelligence | cross-developer peersrobotics |
Questions
OpenVLA 7B FAQs
What is OpenVLA 7B?+
The OpenVLA team's open 7B vision-language-action model trained across Open X-Embodiment robot trajectories. It is developed by OpenVLA Research Team and its current sourced lifecycle status is active.
What inputs and outputs does OpenVLA 7B support?+
OpenVLA 7B's current record lists Text, Image, and Robot state as input and Robot action as output.
What kind of robotics model is OpenVLA 7B?+
OpenVLA 7B is classified as a vision-language-action model. Its action representation is Tokenized actions decoded to continuous robot controls, its control architecture is Fused SigLIP and DINOv2 visual encoder with Llama 2 7B backbone, and inference is recorded as flexible.
Which robots and embodiments work with OpenVLA 7B?+
The developer reports WidowX, Google Robot, and Franka Panda as supported or represented embodiments. No direct model-to-platform relationship is currently verified in the catalog.
How large is OpenVLA 7B?+
The current record lists 7B total parameters and an unknown active parameter count. Its architecture is Fused SigLIP and DINOv2 visual encoder with Llama 2 7B backbone.
What capabilities and tasks does OpenVLA 7B support?+
OpenVLA 7B's recorded capabilities are cross-embodiment, fine-tuning, and generalist-manipulation. Its supported tasks are robot-control, dexterous-manipulation, and pick-and-place.
What training data is reported for OpenVLA 7B?+
970,000 robot manipulation trajectories from Open X-Embodiment described by the authors.
Is OpenVLA 7B available through an API?+
OpenVLA 7B is explicitly marked as not API-available in the current record.
How much does OpenVLA 7B cost through an API?+
OpenVLA 7B has no verified hosted API offering, so no provider token price applies. It is a downloadable or self-hosted artifact whose cost depends on the serving infrastructure and workload.
Can OpenVLA 7B be self-hosted?+
OpenVLA 7B is an open-weight model. Self-hosting is supported, with no license recorded.