DeepSeek's web and API services suffered a partial outage on Nov 7. Multiple users reported that the chat interface displayed a "server temporarily unavailable" message and that API calls returned HTTP 500 Internal Server Error. DeepSeek's official service status page showed a partial disruption of its web and API services, and the company said it was investigating. By the time the status page was updated, the outage had lasted more than three minutes.
What went wrong and what users reported
According to posts by V2EX users, the failure manifested as API requests returning HTTP 500 errors, with the message "Internal Server Error," the type labeled "internal_error," and a request ID attached. Other users posted that "ds is down again," reporting that the web-based chat was also unusable.
IT之家 reported, citing feedback from multiple users, that DeepSeek chats showed a "server temporarily unavailable" message. The report also cited the official service status page, confirming a partial disruption of the web and API services and that the company was investigating.
Official status and scope of impact
DeepSeek's official service status page labeled the incident a "partial outage," indicating that not all services were completely unavailable, though the specific affected modules and user scope were not disclosed in detail. As of publication, the company had not announced the cause of the failure or an estimated recovery time.
As a leading domestic large-model service provider, DeepSeek's API is integrated and used by numerous developers and enterprises, and its web platform also has a large individual user base. The outage could directly affect applications and developers that rely on its API, and everyday use by web users has also been affected.
What to watch
The company is still investigating. Going forward, attention should focus on the root cause, the recovery time, and whether officials issue further statements. For developers, the incident once again highlights the availability risks of third-party large-model services and the need to build disaster recovery and fallback plans.