Your team already has a data and robot interface
Inspect the observation and action formats, normalization and supported integration path for each exact release. Account for conversion and calibration work.
Curated comparison · Open robot learning workflows
OpenVLA vs SmolVLA
Both projects provide resources for vision-language-action research. Compare interfaces, checkpoint terms and adaptation requirements before considering results reported under different test conditions.
Source contextPrimary specReviewed
Sources 4
openvla/openvla official repository
openvla · Publication date not disclosed · Source accessed 2026-09-28
Official README inspected live. Repository code terms do not automatically cover model weights, data, or third-party assets.
OpenVLA research project
OpenVLA research team · Publication date not disclosed · Source accessed 2026-09-28
SmolVLA base model card
Hugging Face LeRobot · Publication date not disclosed · Source accessed 2026-09-28
SmolVLA release
Hugging Face · Publication date not disclosed · Source accessed 2026-09-28
Version and date boundaries matter. The exact revision below defines each column. A deployment of an earlier revision is not evidence that a newer revision is in production. Missing fields remain unknown.
On small screens, swipe the table horizontally. Each row carries its own source and review date.
| Decision dimension | OpenVLA Exact revision / scope OpenVLA 7B Primary specReviewed Sources 1openvla/openvla official repository openvla · Publication date not disclosed · Source accessed 2026-09-28 Official README inspected live. Repository code terms do not automatically cover model weights, data, or third-party assets. | SmolVLA Exact revision / scope lerobot/smolvla_base Primary specReviewed Sources 1SmolVLA base model card Hugging Face LeRobot · Publication date not disclosed · Source accessed 2026-09-28 |
|---|---|---|
| Developer The developer identity does not establish a deployment relationship. | OpenVLA research team Primary specReviewed Sources 1openvla/openvla official repository openvla · Publication date not disclosed · Source accessed 2026-09-28 Official README inspected live. Repository code terms do not automatically cover model weights, data, or third-party assets. | Hugging Face LeRobot Primary specReviewed Sources 1SmolVLA base model card Hugging Face LeRobot · Publication date not disclosed · Source accessed 2026-09-28 |
| Exact version Keep generations, checkpoints and software revisions separate. | OpenVLA 7B Primary specReviewed Sources 1openvla/openvla official repository openvla · Publication date not disclosed · Source accessed 2026-09-28 Official README inspected live. Repository code terms do not automatically cover model weights, data, or third-party assets. | lerobot/smolvla_base Primary specReviewed Sources 1SmolVLA base model card Hugging Face LeRobot · Publication date not disclosed · Source accessed 2026-09-28 |
| System or model type Choose alternatives that address the same decision. | Vision-language-action Primary specReviewed Sources 1openvla/openvla official repository openvla · Publication date not disclosed · Source accessed 2026-09-28 Official README inspected live. Repository code terms do not automatically cover model weights, data, or third-party assets. | Vision-language-action Primary specReviewed Sources 1SmolVLA base model card Hugging Face LeRobot · Publication date not disclosed · Source accessed 2026-09-28 |
| Access conditions Availability of a paper is distinct from access to a model or service. | Public training code and pretrained checkpoint Primary specReviewed Sources 1openvla/openvla official repository openvla · Publication date not disclosed · Source accessed 2026-09-28 Official README inspected live. Repository code terms do not automatically cover model weights, data, or third-party assets. | Public model weights and LeRobot implementation Primary specReviewed Sources 1SmolVLA base model card Hugging Face LeRobot · Publication date not disclosed · Source accessed 2026-09-28 |
| License / reuse Code, checkpoints, datasets and assets can have different terms. | Not publicly disclosed | Apache-2.0 model card Primary specReviewed Sources 1SmolVLA base model card Hugging Face LeRobot · Publication date not disclosed · Source accessed 2026-09-28 |
| Inputs Observation and scene formats determine integration work. | Not publicly disclosed | Multiview images, robot state and optional language Primary specReviewed Sources 1SmolVLA base model card Hugging Face LeRobot · Publication date not disclosed · Source accessed 2026-09-28 |
| Outputs An action policy, image generator and simulator return different outputs. | Not publicly disclosed | Continuous actions Primary specReviewed Sources 1SmolVLA base model card Hugging Face LeRobot · Publication date not disclosed · Source accessed 2026-09-28 |
| Embodiment Hardware and control boundaries are part of the evaluation. | Not publicly disclosed | Not publicly disclosed |
| Compute Confirm requirements for the intended workload. | Not publicly disclosed | Not publicly disclosed |
| Integrations Verify exact versions, interfaces and support responsibilities. | Not publicly disclosed | Not publicly disclosed |
Choose based on your constraints.
These are evaluation paths, not a ranking. Verify the current terms and run a task-specific trial before committing.
Your team needs a defensible experiment
Hold embodiment, data split and evaluation conditions constant. Treat published results as protocol-specific evidence rather than a global ordering of models.
Evidence gaps to resolve
- Published fields describe the identified revision only; configuration and source dates may differ.
- No matched independent operating trial is asserted by this comparison. Missing price, reliability and intervention figures remain unknown.
- Benchmark results require the same protocol, inputs and revision context. Cross-benchmark totals are not calculated.
Inspect the evidence
Sources & evidence
openvla/openvla official repository
openvla · Repository · Publication date not disclosed · Source accessed 2026-09-28
Factual summary and attribution only; no upstream prose, images, weights, or dataset redistributed.
Official README inspected live. Repository code terms do not automatically cover model weights, data, or third-party assets.
Referenced by 5 records: OpenVLA · What Is Physical AI? · Physical AI vs Embodied AI · Physical AI vs Robotics · What Is a VLA Model?
OpenVLA research project
OpenVLA research team · Paper · Publication date not disclosed · Source accessed 2026-09-28
Factual summary and attribution only; no upstream prose, images, weights, or dataset redistributed.
Referenced by 1 record: OpenVLA
SmolVLA base model card
Hugging Face LeRobot · Official spec · Publication date not disclosed · Source accessed 2026-09-28
License: Apache-2.0. Factual summary and attribution only; no upstream prose, images, weights, or dataset redistributed.
Referenced by 2 records: SmolVLA · Hugging Face LeRobot
SmolVLA release
Hugging Face · Official release · Publication date not disclosed · Source accessed 2026-09-28
Factual summary and attribution only; no upstream prose, images, weights, or dataset redistributed.
Referenced by 1 record: SmolVLA