DeepSeek version 3.2 outperformed GPT-5 and Gemini Pro on math and logic benchmarks, improving your code answers and problem ...

DeepSeek's latest AI models are here.

DeepSeek unveils V3.2 AI models matching GPT-5 and Google Gemini 3.0 Pro performance at fraction of the cost, introducing breakthrough sparse attention and reasoning-with-tools capabilities in ...

Deepseek version 3.2 packs 671B parameters with 37B active at inference, giving you faster tool use and lower run costs on ...

China’s DeepSeek faces accusations of training its AI model with banned Blackwell AI chips, though Nvidia says it has found ...

Chinese AI company DeepSeek is accused of smuggling Nvidia chips to power newer AI models, but the spotlight is how this ...

China-based DeepSeek has launched a pair of new artificial intelligence models, DeepSeek-V3.2 and DeepSeek-V3.2-Speciale, which are open-sourced and topped the results of OpenAI's GPT-5 and Google's ...

DeepSeek launched two new AI models designed to rival GPT-5 and Gemini 3.0. Named as DeepSeek-V3.2 and DeepSeek-V3.2 Speciale ...

At a rare public appearance, Chinese artificial intelligence (AI) startup DeepSeek issued a stark warning: the long-term advance of AI could upend jobs and disrupt social structures, and companies ...

Last week, DeepSeek released updated AI models. Ian says they might show China catching up to the U.S. in the AI race.

Explore the full archive of TIME, a century of journalism, insight, and perspective, with AI that helps you research, connect ...