Next Article
DeepSeek's new AI model challenges Western counterparts
In August 2025, Chinese tech company DeepSeek made waves with its V3.1 model—a massive 685-billion-parameter AI that rivals top Western systems.
It scored 71.6% on a major coding benchmark and stands out with a huge 128,000-token memory, making it great for chatbots and scientific tools that need to remember lots of info.
What's the significance of this development?
DeepSeek released V3.1 as open-source, letting anyone use or adapt it—no paywalls or fancy hardware needed.
That means what might cost $70 to run on Western models can now be done for just $1 with DeepSeek.
Even though some countries have banned it over security worries, this move is shaking up the global AI scene and helping China push for more tech independence.