Alibaba unveils more powerful new Zhenwu AI chip and new LLM

Alibaba 5 month 19 day, unveiled a more powerful new Zhenwu AI chip and a new large language model (LLM). The company described the announcement as the latest update to its AI technology. The materials released did not include specific performance figures or pricing information.

According to the company’s announcement, the new Zhenwu chip was presented as a design with greater computing power than the previous generation. The announcement specifically described the chip as designed for agent workloads and operating large models. The company said benchmark figures and test results would be shared in a separate announcement.

The company said the newly unveiled large language model was designed to support agent capabilities. The announcement stated that the model focuses on agent-style applications that combine response generation and task execution. The announcement did not specify the model’s parameter count or publicly released model name.

The presentation introduced the chip and model as part of an upgrade to Alibaba’s AI stack. The company said it plans to apply the technology to its cloud services and its own services. It said information on the commercialization timeline and initial customers would be provided in a later announcement.

The announcement identified support for agent workloads as a key goal of the technology. The company explained that agent-based applications are linked to demand from cloud customers. No specific customer examples or initial adoption partners were presented in this announcement.

Alibaba announced that it aims to expand AI processing capabilities and the scope of applications through its chip and model. It said further details and the commercialization schedule would be disclosed later. The announcement only specified that this was part of its product and service roadmap.