Recently, DeepSeek officially announced that its latest version of API is now open for testing. As a highly anticipated AI product, the new version has achieved a groundbreaking upgrade in proxy capabilities, with baseline test results even surpassing previous preview versions, demonstrating an extremely strong overall strength.

In practical application tests, an AI deep developer was the first to integrate the new version of the API into the system for testing. The results showed that when facing the same high-difficulty complex tasks and prompts, the new version completed the task efficiently in about 40 seconds, while overseas top-tier large models took 1 minute and 47 seconds. Although overseas models still have a slight advantage in some aspects of output quality, the delivery speed and usability of the new version have already reached industry-leading standards.

Aside from the impressive processing speed, price and cost have also become core highlights of this update. According to the official's latest pricing mechanism, the input cost of some versions has dropped significantly, with a substantial overall reduction. However, different versions have their own strategies, with some high-performance or specific time period API costs adjusted accordingly. Overall, compared to mainstream overseas official large models, domestic AI still maintains a highly competitive advantage in API pricing, with cost differences generally ranging from tens to hundreds of times.

Industry insiders analyze that as domestic large models continue to break through in performance, speed, and cost control, the commercialization threshold of AI technology is further decreasing. In the future, more developers and enterprises will be able to enjoy efficient AI computing power at extremely low costs, and the industry ecosystem may witness a new round of reshaping.