SI Agent News
DevelopingOpen models

DeepSeek users report new caps on long chats and daily use

Users say DeepSeek's app and website now limit long conversations, input size and daily use, a sign that demand is outrunning computing capacity.

801%week-over-week rise in DeepSeek V4.1 Flash token calls on OpenRouter

The short version

Why it matters

If you use DeepSeek's chat as a long-running notebook, a chat left idle for a week may stop working. Free AI tools run on limited capacity, so their terms can tighten with little warning. It is wise to keep your own copies of important work and a tested backup tool.

What users are reporting

As DoNews reported on November 16, some users on social platforms say DeepSeek has started limiting long conversations in its app and web versions. Several received a message saying that, because of insufficient computing power, chats inactive for more than 7 days are no longer supported.

Two other limits are reported. The text allowed in a single input has been shortened to about 40,000 characters, and there is a daily conversation quota. Once you hit it, the page says usage has reached its limit because servers are busy.

No official word yet

At the time of the report, DeepSeek had not published a separate announcement about the limits or said when they might be lifted. So the details come from user screenshots and reports, not from the company. Exact quota numbers are not given.

A pattern of strain

The report presents this as a continuation of recent capacity trouble. DeepSeek's official status page records 8 service interruptions since October 2026. They affected the chat service and the API (the connection developers use to plug the model into their own software) for V4 Pro and V4.1 Flash.

For V4.1 Flash's API, availability from August to October was 99.46%. Stretched over a year, that works out to roughly 47.3 hours of downtime.

Demand is surging

One source of pressure is demand. OpenRouter, a platform that lets people use many models in one place, ranks them weekly by tokens, the small chunks of text models process. For October 19 to 25, DeepSeek V4.1 Flash handled 33.6 trillion tokens, up 801% from the week before and first in the world.

In late July and early August the weekly figure was 7.22 trillion. That is more than a fourfold rise in under three months.

What to watch

Try it

Check your longest regular prompt against the roughly 40,000-character input limit. At the end of important chats, ask for a short summary and save it in your own notes.

Sources and links

Liked it? Say why

Add a line for other readers.

No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.

Listen to the AI Wire headlines

Listen to the AI Wire

Listen to our daily AI shows every morning: on your commute, in the shower, wherever you are

Listen to the daily shows

First reported by DoNews · Read the original. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.

More AI news