Chinese AI startup DeepSeek has announced an upgrade to its R1 model, the reasoning-focused artificial intelligence system that brought it international attention earlier this year.
The update, described as a “minor trial upgrade,” is now open for user testing, though the company has not disclosed technical details and declined to comment further.
The original R1 model, launched in January, outperformed several Western systems on key benchmarks at a fraction of the usual development cost.
Its release sparked a wave of new AI models from other Chinese firms, including Alibaba and Zhipu AI, and prompted a reassessment of spending on AI infrastructure.