Live status
Is DeepSeek down right now?
DeepSeek: no problems in our tests of DeepSeek's models in the last 15 minutes · DeepSeek has no public status feed · checked
Updated . While people are reading it, this page is refreshed about once a minute.
Our tests of DeepSeek's models
We send "Reply with OK" to DeepSeek V4.1 Flash every 5 minutes and DeepSeek V4 Pro every 30 minutes, and time how long each takes to start answering and to finish. The verdict covers the last 15 minutes: our last three tests.
Last test
- Answered, last 15 min
- 100%
- Answered, last hour
- 100%
- Answered, last 24 h
- 100%
- Median first token, 24 h
- 0.5 s
The last 24 hours
| Model | Last 24 h | First token | Whole reply | Last test |
|---|---|---|---|---|
| DeepSeek V4.1 FlashEvery 5 minutes | 43 of 43 answered | 0.4 s | 0.5 s | AnsweredThrough OpenRouter, |
| DeepSeek V4 ProEvery 30 minutes | 7 of 7 answered | 0.6 s | 1.0 s | AnsweredThrough OpenRouter, |
No failed tests in the last 24 hours.
What DeepSeek's status page says
DeepSeek's own status page, read at most once a minute. It reports incidents once DeepSeek's team confirms them, so it can lag behind what users and our tests see.
DeepSeek's status page has no public data feed we can read, so we link it instead. Its status page: status.deepseek.com.
What our test covers, and what it doesn't
Our test calls DeepSeek's models through OpenRouter, the way llmwise does, on hosts that don't store or train on prompts, not on DeepSeek's own servers. So it tells you whether the models answer; DeepSeek's own chat and API can be busier than that.
Each test is one message of a few words, with the reply capped at 32 tokens, tried once, with no retry and no switch to another route. A test counts as failed when the model refuses it as busy (a rate limit or "overloaded"), breaks (a server error), drops the connection, or takes more than 30 seconds. A test our own request broke doesn't count against DeepSeek.
What to do now
If our test passes, DeepSeek's models are answering elsewhere: the busy part is DeepSeek's own site or API.
Wait a minute and resend: "server busy" usually passes within minutes.
Check status.deepseek.com for a posted incident.
Use the same models through another host, or another company's model.
Ask another model
llmwise has Claude, GPT, Gemini and more in one chat, so an outage at one company doesn't stop your work. The free trial is 5 messages, with a sign-in link by email and no card.
DeepSeek errors, and how to fix them
The errors people see most with DeepSeek, what each one usually means, and what to try.
"Server busy, please try again later"
DeepSeek's own chat is out of room.
Wait a minute and resend, or try at a quieter hour. The same models are available from other hosts.
API 503 "Server Overloaded"
DeepSeek's API is too busy for the request.
Retry with a growing wait between tries. If it lasts, try another host of the same model.
API 429 "Rate Limit Reached"
Too many requests from your account in a short time.
Send requests more slowly, and retry after a pause.
API 402 "Insufficient Balance"
Your DeepSeek account has run out of paid balance.
Top up your DeepSeek account. It isn't an outage.
Other AI status pages
Questions
Why does DeepSeek say "Server busy, please try again later"?
It's DeepSeek's own chat telling you it has no room for your message right now. It's common at busy hours and passes on its own. If our test of DeepSeek's models shows failures too, the models themselves are struggling.
Why can your test pass while DeepSeek's site is busy?
Our test reaches DeepSeek's models through OpenRouter, which sends them to hosts other than DeepSeek's own servers (llmwise never uses hosts that may train on prompts). So our results can look fine while DeepSeek's own chat is busy, and that's useful to know: the models are there, just not on DeepSeek's site.
Why don't you show DeepSeek's official status?
Its status page, status.deepseek.com, doesn't offer a public data feed we can read, so we link it rather than repeat it.
What can I use while DeepSeek is down?
The same DeepSeek models through llmwise, or another company's model: Claude, GPT, and Gemini are in the same chat, and the free trial is 5 messages with no card.
Claude, GPT, Gemini, DeepSeek, Grok, Kimi, and GLM, in one chat.
See what a message costs before you send it. Free is 5 messages to try; sign in with an email link, no password or card.