Search
DeepSeek Unveils New Model with Support for Million-Token Context
13 hour ago /
Read about 0 minute
Author：小编
Numerous users have observed that DeepSeek has rolled out version updates across its web and app interfaces, now boasting a context capacity of up to 1M (one million) tokens. Prior to this, the DeepSeek V3.1 iteration, launched in August of the previous year, had already enhanced its context length to 128K. Based on hands-on testing by journalists, DeepSeek asserts its ability to handle 1M tokens, facilitating the simultaneous processing of exceptionally lengthy texts. During the test, the journalist submitted a copy of the novel 'Jane Eyre', which comprises over 240,000 tokens, and DeepSeek successfully identified and processed the document's contents.
-
C114 Communication Network -
Communication Home
7 X 24 Track global technological trends
