The Tai Chu Yuanqi completed the Zhifu GLM-5.0 and the Ali Qianwen dual open source model deep adaptation.

On the evening of February 18th, "Taichu Yuanqi" completed the deep adaptation work of multiple domestic mainstream open-source large models, including Zhigu GLM-5.0 and Ali Qianwen Qwen3.5-397B-A17B. It is understood that the adaptation of domestic large models by Taichu Yuanqi is carried out on its self-developed T100 acceleration card. In the process of adapting mainstream large models, Taichu Yuanqi has launched a ladder-style development toolchain in the SDAA software stack to cover the diverse development needs from entry-level to advanced levels according to the different technical abilities of developers. This toolchain can help developers quickly build high-performance operators, achieve seamless compatibility with mainstream AI ecosystems, and significantly reduce the technical threshold and cost of migrating to the CUDA ecosystem.