Technology ❯Artificial Intelligence ❯Chatbots ❯DeepSeek

Model Development

DeepSeek Upgrades R1 AI Model to Challenge Top Western Rivals

Open-source under an MIT license, the update boosts AIME2025 math accuracy to 87.5%.

China's DeepSeek AI Startup Draws Global Attention and Government Oversight

Chinese AI Startup DeepSeek Challenges Industry Norms With Ultra-Efficient Models

DeepSeek's AI Breakthrough Shakes Global Tech Landscape

China's DeepSeek AI Challenges U.S. Tech Giants with Breakthrough Efficiency