On June 4, 2025, the artificial intelligence development platform Hugging Face unveiled a new open source AI model dedicated to robotics, called SmolVLA. According to the company, this model stands out for its lightness and its ability to run on consumer-grade equipment, including a MacBook, thus making advanced robotics more accessible. This announcement marks an important step in Hugging Face’s efforts to democratize access to robotics technologies.
A compact and high-performing model
SmolVLA, with its 450 million parameters, was trained on datasets shared by the community under compatible licenses. This data comes notably from the LeRobot Community Datasets, robotics-specific datasets hosted on Hugging Face’s AI development platform. Despite its relatively modest size, the company claims that SmolVLA outperforms much larger models, both in virtual and real environments. Parameters, often called “weights,” are the internal components of the model that guide its behavior.
What makes SmolVLA particularly innovative is its ability to run on affordable hardware, such as a consumer-grade GPU or even a MacBook. This feature significantly reduces barriers to entry for developers and enthusiasts wishing to explore sophisticated robotics projects from home. Furthermore, SmolVLA supports an asynchronous inference stack, allowing the processing of a robot’s actions to be separated from what it perceives visually and audibly. According to Hugging Face, this separation improves robot responsiveness in constantly evolving dynamic environments.
A democratization approach
In a blog post, Hugging Face explains that SmolVLA aims to “democratize access to vision-language-action (VLA) models and accelerate research toward generalist robotic agents.” This model is part of a broader initiative by the company to create an ecosystem of low-cost robotics hardware and software. Last year, Hugging Face had already launched LeRobot, a collection of models, datasets, and tools focused on robotics. More recently, the acquisition of Pollen Robotics, a French startup, allowed the company to offer affordable robotic systems, including humanoid robots.
SmolVLA is available for download on the Hugging Face platform, and users have already reported successful applications, such as controlling a third-party robotic arm. A user shared their experience, claiming to have used the model to operate a Koch robotic arm with an RTX 2050 card (4 GB), after fine-tuning with just 31 demonstrations. These promising results highlight the potential of SmolVLA to drive innovation in the field of robotics.
An expanding ecosystem
Hugging Face is not content with just developing AI models. The company has committed to rapidly expanding its robotics efforts. In 2024, it launched LeRobot, and in April 2025, it acquired Pollen Robotics, thereby strengthening its position in the field. Additionally, Hugging Face has unveiled several affordable robotic systems, including humanoids, available for purchase. These initiatives aim to make robotics accessible not only to large companies, but also to researchers, independent developers, and hobbyists.
Growing competition
Hugging Face is not alone in the race for open source robotics. Major players like Nvidia offer tools for open robotics, while startups like K-Scale Labs work on components for “open source humanoids.” Other companies, such as Dyna Robotics, Physical Intelligence (backed by Jeff Bezos), and RLWRLD, are also heavy competitors in this rapidly emerging sector. Despite this competition, Hugging Face distinguishes itself through its accessibility and community-focused approach, with freely shared models and tools.
Toward an accessible future for robotics
The launch of SmolVLA represents a significant step forward toward affordable and accessible robotics solutions. By enabling high-performing AI models to run on consumer-grade hardware, Hugging Face paves the way for faster and more inclusive innovation. Whether for home projects or professional applications, SmolVLA could transform the robotics landscape by making advanced technologies accessible to a wider audience.
For more information, the SmolVLA model is available for download on the Hugging Face platform, and users can already begin exploring its many possibilities.
Sources
- TechCrunch, “hugging face says its new robotics model is so efficient it can run on a macbook,” June 4, 2025, https://techcrunch.com/2025/06/04/hugging-face-says-its-new-robotics-model-is-so-efficient-it-can-run-on-a-macbook/
- Hugging Face blog: https://huggingface.co/blog/smolvla