DeepSeek Launches V4, Challenges US AI Giants

Chinese AI startup DeepSeek has released preview versions of its V4 model, featuring improved reasoning, knowledge, and agentic capabilities in pro and flash variants. The update expands its context window from 128,000 to 1 million tokens and claims superior performance against some top US models. Analysts call V4 competitive but not as groundbreaking as DeepSeek's earlier R1 model, which rattled markets in January 2025. Independent evaluations are still needed, and US rivals Anthropic and OpenAI have previously accused DeepSeek of improperly building on their technology.