The official Silicon-based liquidity Weibo recently released an important announcement announcing that the DeepSeek-R1 and V3 APIs of its SiliconCloud platform now support Batch Inference. The launch of this new feature marks another major breakthrough in silicon-based flow in the field of artificial intelligence services.
Through the batch API, users can send large amounts of data processing requests to the SiliconCloud platform without worrying about real-time inference rate limiting. According to official introduction, batch inference tasks are expected to be completed within 24 hours, greatly improving the efficiency of data processing. What is even more exciting is that the price of DeepSeek-V3 batch inference is directly reduced by 50% compared to real-time inference. In addition, from March 11 to March 18, DeepSeek-R1 batch reasoning launched a 75% discount activity, with the input price only 1 yuan/million Tokens, and the output price also dropped to 4 yuan/million Tokens.
The introduction of batch inference function provides users with more efficient solutions for handling large-scale data tasks. Whether it is generating reports, cleaning data, or performing data analysis, model performance evaluation and other scenarios, this new function can play an important role. Especially suitable for data processing tasks that do not require real-time response, allowing users to enjoy the high-quality services of DeepSeek-R1 and V3 API at a lower cost.
