[Moore Threads Completes Full-Link Support for Protenix-v2] According to a report by Jinse Finance, on September 24, Moore Threads, leveraging its proprietary MUSA software stack, officially completed full-link inference support for the open-source biomolecular structure prediction model Protenix-v2 on its unified training and inference intelligent computing card, MTT S5000. Real-world testing shows that the end-to-end inference performance of the S5000 is approximately 26% faster than the average of leading international GPUs, with highly aligned accuracy in 3D conformation predictions. It is understood that Protenix-v2 was developed by ByteDance's Seed team and officially released in April 2026. The model overcomes the limitations of incomplete openness of top overseas model weights and the inability to deploy privately, making it one of the few open-source platforms globally capable of high-precision end-to-end predictions for proteins, DNA, RNA, and small molecule ligand complexes. (Securities Times)