AI & ML Research 8 Hours

Last updated: June 22, 2026, 5:30 AM ET

AI & ML Research

DeepSeek released a new family of large language models, including a 1.3 trillion parameter model, designed to outperform OpenAI's GPT-4 on several benchmarks. The models reportedly achieve a 10.8% improvement in reasoning capabilities and a 6.1% uplift in coding tasks compared to existing state-of-the-art systems. This development signals intensified competition among major AI labs for developing more capable and efficient foundational models.