This open source robotic brain thinks in 3D

European roboticists today Released robust open source artificial intelligence A model that acts as an artificial brain Robots– Helping them understand things and deal with them with new skill.

new model, Spear-1It was developed by researchers at the Institute of Computer Science, Artificial Intelligence and Technology (INSAIT) in Bulgaria. It may help other researchers and startups build and trial smarter devices for factories and warehouses.

Just as open source language models have made it possible for researchers and companies to experiment with generative AI, SPEAR-1 should help roboticists experiment and iterate quickly, says Martin Vetchev, a computer scientist at INSIAT and ETH Zurich. “Open-weighted models are essential for the development of embodied AI,” Vetchev told WIRED ahead of the release.

SPEAR-1 differs from existing robotics models in that it incorporates 3D data into its training mix. This gives the model an enhanced understanding of the physical world, making it easier to understand how objects move through physical space.

Basic robot models are generally built on vision language models (VLMs) that have a broad but limited understanding of the physical world because training tends to come from 2D images. “Our approach addresses the mismatch between the 3D space in which the robot operates and the VLM knowledge that forms the core of the robotic foundation model,” says Vetchev.

The SPEAR-1 has roughly the same capacity as commercial foundation models designed to power robots, when scaled On RoboArena, it is standard This model tests the model’s ability to make the robot do things like squeeze a ketchup bottle, close a drawer, and staple pieces of paper together.

The race to make robots smarter has already begun Billions of dollars Ride it. The commercial potential of general-purpose robots has given rise to well-funded startups including… sporadic and specialist next to Physical intelligence. SPEAR-1 is almost as good as Pi-0.5 in terms of physical intelligence, A billion-dollar startup founded by an all-star team of robotics researchers.

SPEAR-1 suggests that the quest to build smarter robots may include both closed models like those at OpenAI, Google, and Anthropic, as well as open source variants like Llama, DeepSeek, and Qwen.

However, robot intelligence is still in its infancy. It is possible to train an AI model to operate a robot arm so that it can reliably pick up certain objects from a table. However, in practice, the model will need to be retrained from scratch if a different type of robot arm is used or if the object or environment is changed.