Search
DeepSeek Updates Model to Process Ultra-Long Texts in One Go
16 hour ago /
Read about 0 minute
Author：小编
The new DeepSeek model is now live. When you open the DeepSeek App, you may receive an update prompt—simply tap 'Update Now' to try it out. Currently in grayscale testing, the model's standout feature is a significant boost in context processing capacity, jumping from 128K Tokens to 1M Tokens, enabling it to handle nearly 10 times more information at once. Meanwhile, the knowledge base has been updated to May 2025, likely incorporating a newly trained foundational model based on more recent data. This update brings enhanced context capacity, updated knowledge, improved capabilities, and a more natural interactive experience.
-
C114 Communication Network -
Communication Home
7 X 24 Track global technological trends
