Recently, Moore Thread officially completed full-link inference support for the open source biomolecular structure prediction model ProteniX-v2 on the MTT S5000 integrated intelligent computing card based on the autonomous MUSA software stack. Actual measurements show that the S5000's end-to-end inference performance is about 26% faster than the average speed of mainstream international GPUs, and the accuracy of 3D conformation prediction is highly aligned. According to information, Protenix-v2 was developed by the ByteDance Seed team and officially released in April 2026. It breaks the restrictions that overseas top model weights are not fully open and cannot be privately deployed. It is currently one of the few open source platforms in the world that can perform high-precision end-to-end prediction of proteins, DNA, RNA, and small molecule ligand complexes.

Recently, Moore Thread officially completed full-link inference support for the open source biomolecular structure prediction model ProteniX-v2 on the MTT S5000 integrated intelligent computing card based on the autonomous MUSA software stack. Actual measurements show that the S5000's end-to-end inference performance is about 26% faster than the average speed of mainstream international GPUs, and the accuracy of 3D conformation prediction is highly aligned. According to information, Protenix-v2 was developed by the ByteDance Seed team and officially released in April 2026. It breaks the restrictions that overseas top model weights are not fully open and cannot be privately deployed. It is currently one of the few open source platforms in the world that can perform high-precision end-to-end prediction of proteins, DNA, RNA, and small molecule ligand complexes.