Search papers, labs, and topics across Lattice.
The paper introduces SimVLA, a simplified Vision-Language-Action (VLA) model for robotic manipulation designed as a transparent baseline for research. SimVLA decouples perception from control, uses a standard vision-language backbone with a lightweight action head, and standardizes training dynamics. The model achieves state-of-the-art performance on simulation benchmarks without robot pretraining, outperforming models with billions of parameters, and reaches on-par real-robot performance compared to pi0.5.
A surprisingly simple 0.5B parameter VLA model, SimVLA, outperforms multi-billion parameter models in robotic manipulation, suggesting architectural complexity isn't always necessary.
Vision-Language-Action (VLA) models have emerged as a promising paradigm for general-purpose robotic manipulation, leveraging large-scale pre-training to achieve strong performance. The field has rapidly evolved with additional spatial priors and diverse architectural innovations. However, these advancements are often accompanied by varying training recipes and implementation details, which can make it challenging to disentangle the precise source of empirical gains. In this work, we introduce SimVLA, a streamlined baseline designed to establish a transparent reference point for VLA research. By strictly decoupling perception from control, using a standard vision-language backbone and a lightweight action head, and standardizing critical training dynamics, we demonstrate that a minimal design can achieve state-of-the-art performance. Despite having only 0.5B parameters, SimVLA outperforms multi-billion-parameter models on standard simulation benchmarks without robot pretraining. SimVLA also reaches on-par real-robot performance compared to pi0.5. Our results establish SimVLA as a robust, reproducible baseline that enables clear attribution of empirical gains to future architectural innovations. Website: https://frontierrobo.github.io/SimVLA
