Moore threads announces full support for Protenix-v2 model on mtt s5000

Moore Threads has announced that its MTT S5000 training and inference AI accelerator now fully supports the open-source Protenix-v2 model, developed by ByteDance's Seed team. The model, released in April 2026, performs on par with or better than DeepMind's AlphaFold 3 in key benchmarks. It is one of the few open-source models capable of high-precision end-to-end prediction of proteins, DNA, RNA, and small molecule ligand complexes. The integration of Protenix-v2 with MTT S5000 marks a milestone for Moore Threads, enabling end-to-end engineering validation across three AI for Science (AI4S) core scenarios: large molecule structure prediction, whole-genome sequence modeling, and clinical medical imaging. This development provides domestic pharmaceutical companies and research institutions with a foundational AI4S computing platform.