DeepSeek, a Chinese artificial intelligence startup, released an update to its R1 reasoning model, thereby intensifying competition with US counterparts like OpenAI.

DeepSeek has not yet issued an explicit public announcement, despite the fact that it has launched R1-0528 on the developer platform Hugging Face.

The model’s description and comparisons were not published.

However, the LiveCodeBench leaderboard, a benchmark that was developed by researchers from UC Berkeley, MIT, and Cornell, rated DeepSeek’s updated R1 reasoning model just marginally behind OpenAI’s o4 mini and o3 reasoning models on code generation.

It also ranked it ahead of xAI’s Grok 3 mini and Alibaba’s Qwen 3.

It stated that a representative from DeepSeek had informed a WeChat group that it had completed a “minor trial upgrade” and that users could begin evaluating it.

Earlier this year, DeepSeek challenged the notion that US export controls were impeding China’s AI advancements by releasing AI models that were on par with or superior to industry-leading models in the United States at a fraction of the cost.

The launch of R1 in January resulted in a significant decline in tech shares outside of China and challenged the notion that scaling AI necessitates a significant investment and computing capacity.

Since the release of R1, Chinese technology giants such as Alibaba and Tencent have introduced models that purport to surpass DeepSeek’s capabilities.

OpenAI has reduced prices and released an o3 mini model that requires less computing power, while Google’s Gemini has implemented discounted access tiers.