DeepSeek's AI Models Now Available on Huawei's Ascend Cloud Service
The Launch of DeepSeek AI Models
During the Lunar New Year holidays, Huawei's cloud unit partnered with Beijing-based start-up SiliconFlow to introduce DeepSeek's large language model V3 and reasoning model R1 to end users via the Ascend cloud service. This platform promises performance comparable to premium global GPUs.
Affordability and Access
- V3 access costs 1 yuan ($0.13) for 1 million input tokens.
- Output tokens for V3 are priced at 2 yuan.
- R1 charges are set at 4 yuan for input and 16 yuan for output tokens.
The Global Impact of DeepSeek
As the open-source DeepSeek models attract attention in both China and the US, tech giants like Microsoft and Amazon.com are beginning to integrate R1 into their own services. Microsoft supports R1 on its Azure cloud platform and GitHub, facilitating local AI applications on personal computers.
Domestic Advancement in AI
This deployment reinforces China's ambition to strengthen its AI industry, reflecting significant progress in reducing reliance on foreign technology. Huawei's collaboration with SiliconFlow marks a pivotal step in bolstering domestic AI capabilities.
Industry Reception and Future Outlook
Emerging start-up DeepSeek has startled major industry players in Silicon Valley for its ability to deliver AI models at significantly reduced costs compared to larger US competitors. Tencent has also embraced DeepSeek's R1 model, promoting its easy setup process.
This article was prepared using information from open sources in accordance with the principles of Ethical Policy. The editorial team is not responsible for absolute accuracy, as it relies on data from the sources referenced.