Lumos Prime R0 Tops MolmoSpaces Global Benchmark for Zero-Shot Embodied AI

Lumos Prime R0 Achieves Top Ranking on the MolmoSpaces Benchmark

Lumos Robotics, a leading Chinese embodied AI company, has claimed the top spot on the latest MolmoSpaces leaderboard, a globally recognized benchmark for zero-shot embodied AI, with its industrial embodied foundation model Prime R0. Prime R0 achieved the highest overall success rate across both single-arm fine manipulation and dual-arm collaboration tasks, outperforming leading competitors including NVIDIA’s Cosmos (16B), MIT, and Princeton University. Despite having only 2.8 billion parameters, less than one-sixth the size of NVIDIA Cosmos, Prime R0 delivered the best overall performance. The achievement represents one of the strongest international benchmark performances by a Chinese-developed embodied AI model and validates Lumos’ industrial-first approach to embodied intelligence.

Prime R0 Outperforms Larger Embodied AI Models

The MolmoSpaces benchmark, developed by researchers at Ai2, specifically tests policies on eight base tasks including navigate-to, pick, pick-and-place, and open-door operations. These evaluations occur in synthetic environments that the robot has never encountered, ensuring a true measure of zero-shot performance. The benchmark suite was designed to distinguish performance among various state-of-the-art policies, including Vision-Language-Action models. Unlike traditional benchmarks that rely on fine-tuning with specific data, MolmoSpaces requires models to succeed immediately upon deployment.

Lumos Focuses on Industrial Deployment Rather Than Laboratory Demonstrations

While much of the embodied AI industry remains focused on general-purpose demonstrations, Lumos believes the next stage of industry growth depends on deploying robots into production-ready environments where they can solve practical problems and generate measurable productivity gains. To support this vision, Lumos has built a closed-loop ecosystem spanning industrial data, foundation models, robot hardware, and deployment scenarios. Developed within this framework, Prime R0 prioritizes generalization, responsiveness, and reliability in complex industrial settings rather than pursuing ever-larger parameter counts.

Prime R0 Demonstrates Strong Performance Across Real-World Industrial Tasks

Integrated with the Lumos Touch robotic arm, Prime R0 has already demonstrated strong performance across demanding real-world applications, including sequential floral arrangement, textile manipulation, precision parts storage, and sorting disorganized materials in confined spaces. These deployments highlight the model’s ability to operate effectively across diverse industrial workflows while maintaining reliability and adaptability in complex environments.

Prime R0 Combines Vision-Language-Action Intelligence with World Models

Prime R0 combines Vision-Language-Action (VLA) decision-making with world-model-based prediction, enabling the system to anticipate the physical consequences of actions before execution. This creates a unified loop integrating perception, reasoning, and action. The model is powered by four proprietary technologies: temporally adaptive action generation, unified geometric action representation, lightweight implicit physical prediction, and mixture-of-experts (MoE) networks. Together, these capabilities move beyond traditional perception-driven robotics and form the foundation of Prime R0’s industry-leading performance.

Lower Hardware Costs, Faster Inference, and Greater Operational Efficiency

Many embodied AI systems excel in demonstrations but struggle to meet factory requirements around cost, reliability, cycle time, and scalability. Prime R0 was designed from the outset to address these industrial constraints. The model can run locally on a consumer-grade NVIDIA GeForce RTX 5060 8GB GPU, reducing hardware costs by more than 80% compared with solutions that rely on expensive cloud infrastructure. Prime R0 also delivers millisecond-level inference, supports picking, assembly, sorting, and dual-arm collaboration within a single model, and natively supports Franka robotic arms without retraining. Leveraging SE(3) geometric priors and hybrid real-virtual training data, the model maintains strong performance under industrial challenges such as occlusion, stacking, and positional deviations, helping reduce manual intervention and improve overall equipment effectiveness (OEE).

Lumos NexCore and the Future of Industrial Embodied Intelligence

Lumos views Prime R0 not as an isolated model breakthrough, but as the result of a broader full-stack strategy spanning robotic hardware, industrial data systems, physical AI engines, foundation models, and deployment toolchains. At the center of this strategy is Lumos NexCore, a physical AI platform designed to continuously connect data, models, and real-world deployments through a closed-loop architecture. Unlike approaches focused primarily on academic research or cloud-scale computation, Lumos NexCore is built specifically for industrial deployment and guided by four principles: industrial value first, full-domain system foundation, physical intelligence at the core, and full-stack in-house development. “Lumos NexCore is an operating system for industrial embodied intelligence, and Prime R0 is the first flagship model built on that foundation,” said Yu Chao, Founder and CEO of Lumos Robotics. “We will continue expanding from manufacturing into logistics and additional sectors, with the goal of building the foundational platform for next-generation industrial robotics worldwide.” Prime R0’s success on MolmoSpaces demonstrates how deployment-driven innovation can make embodied intelligence practical, scalable, and ready for real-world industrial environments.