ai
Chinese Startup DeepSeek Releases Preview of V4 AI Model Using Huawei Chips
Chinese startup DeepSeek has released a preview of its V4 AI model, which is designed to operate on chips from Huawei Technologies. The model represents a shift from the company's prior use of Nvidia hardware. DeepSeek stated that the V4 Pro version outperforms most open-source models on world knowledge benchmarks.
BenzingaAudio version
Tap play to generate a narrated version.
A Chinese startup, DeepSeek, released a preview of its V4 artificial intelligence model on Friday. The model is built to run on chips produced by Huawei Technologies. This development follows the company's earlier use of hardware from Nvidia Corp. DeepSeek stated that its V4 Pro model performs better than most open-source competitors in world knowledge benchmarks.
The V4 model is optimized for processing long and complex text tasks. It can handle more than one million tokens while requiring less computational resources compared to similar models. However, the model does not include multimodal features, such as the ability to process images or videos.
The release occurs amid U.S. restrictions on technology exports to China, which have affected access to advanced chips. DeepSeek's move to Huawei chips reflects efforts to use domestically produced hardware. The company did not provide additional details on the timeline or specific reasons for the shift.
Transparency
4 sources · across multiple outlets
CorroborationLimited · 4 sources
