Bolt is a GPU-accelerated musculoskeletal simulator with native support for OpenSim models and designed for massively parallel (1k+) environments at hundreds-to-thousands times real-time speed (in aggregate) on a single GPU.

Bolt is available open-source here. Implementations of RL environments using Bolt are available here.

The following motions were generated without any motion demonstrations, using only a simple reward function: maximize speed without hyperextending joints. Diverse behavior is obtained via task-specific episode termination conditions.

Please refer to our SIGGRAPH Asia paper for a description of Bolt and implementation of RL training environments using the simulator.

Abstract

We present a muscle-driven simulation system for generating biomechanically accurate motion for high-speed athletic locomotion tasks that does not require motion demonstrations. Our approach integrates state-of-the-art biomechanical athlete models into a new, high-performance GPU simulator capable of running at 1000x real-time. High-throughput simulation enables large-batch reinforcement learning to train control policies that operate directly in the model’s high-dimensional muscle excitation space, and are guided only by task-specific episode termination conditions and a reward that encourages maximizing speed while reducing forces needed to respect joint limits. These policies train within a few hours on a single GPU and generate “near visually realistic” motions for complete athletic activities such as a full 100-meter sprint or performing popular athletic locomotion drills like side-shuffling, backpedaling, and carioca. The generated sprinting motions also exhibit strong agreement with experimental data captured from sprinters.

Citation

            
@article{wang26bolt,
    title   = {Learning Realistic Athletic Sprinting Without Demonstrations},
    author  = {William Wang and Nicholas Bianco and Guy Tevet and
               Jennifer Hicks and C. Karen Liu and Scott Delp and
               Kayvon Fatahalian},
    journal = {ACM Trans. Graph.},
    volume  = {45},
    number  = {6},
    year    = {2026}
}