Qiniu Cloud (02567) DeepSeek series models fully launched, Full version for comprehensive experience, Distilled version for one-click deployment.

On February 11th, the well-known audio and video service platform Qiniu Cloud (02567) announced the launch of the DeepSeek-R1 distilled version model. Combined with the announcement on February 7th, Qiniu Cloud has already launched exclusive GPU cloud servers and Token API inference services adapted for the DeepSeek-V3/R1 full range of models. It is reported that Qiniu Cloud hopes that its one-stop AI inference solution can meet the growing demand for generative AI from customers, and provides two deployment methods for models: one is to quickly integrate through API interface, and the other is to deploy with one click on high-performance GPU cloud servers. In order to shorten the distance from idea to product, Qiniu Cloud hopes that developers can integrate the top-notch inference capability comparable to the official DeepSeek-R1 in their own applications. They first launched the full-blooded version of the DeepSeek-R1 671b full parameter model. As an open source model, DeepSeek-R1 matches the closed-source models of overseas giants in inference efficiency, scene adaptation, and many other dimensions, setting new benchmarks. Based on this, Qiniu Cloud provides the most complete and cost-effective 1.73bit quantized DeepSeek-R1 671b full parameter version, greatly reducing the memory requirement to 158GB, enabling fast inference on high cost-effective GPU servers such as A100, V100, and A10. At the same time, in order to meet the differentiated needs of different developers, Qiniu Cloud quickly launched the DeepSeek-R1 distilled version model. This series of models are optimized, with the characteristics of lightweight and efficient inference, and can run on devices with limited computing resources, making them especially suitable for deployment by small and medium-sized enterprises and developers, integrating low-cost applications into actual business practices. To facilitate developers, Qiniu Cloud provides recommended GPU host configurations compatible with different distilled models, and offers a one-click deployment method, ensuring that developers can quickly choose configurations, start and run models according to their own needs.