Skip to content
Brief Business Efficiency ·

DeepSeek apologises and refunds a week of use after 5-hour outage

The Chinese AI firm's chatbot and API were down for nearly five hours on Thursday, and it will repay users in credits.

In brief
  • DeepSeek's chatbot and API, including V4 Pro and V4.1 Flash, were down for about 4 hours 50 minutes on Thursday. The firm will refund October 8 to 14 use as credits.
  • DeepSeek gave no cause. It says it has eliminated single points of dependency but gave no detail on what changed.

The South China Morning Post reported that DeepSeek's chatbot and API services went offline on Thursday morning. An API is the link that lets other software use an AI model. The outage covered the V4 Pro and V4.1 Flash models and hit users worldwide. Service returned around 1pm, after four hours and 50 minutes. In a notice on its API platform, the Hangzhou firm apologised. It will refund all use from October 8 to October 14 as credits.

DeepSeek did not say what caused the outage. Its notice says it has reviewed and eliminated single points of dependency, meaning parts whose failure can stop a whole service. It gave no detail on what failed or what changed. No count of affected users and no date for the credits have been given.

This incident stopped a chatbot and two models together, which shows the risk of relying on one AI provider. Buyers can ask any vendor what fails together, whether a backup exists, what the contract promises for outages, and whether credits cover lost work.

A WebPulse Brief: a short report of an important event, written by the WebPulse Newsroom with AI assistance and checked against the reporting below. When there is more to explain, we follow up with a full story. How we use AI.

Reporting: South China Morning Post.