Chinese AI startup DeepSeek has quietly rolled out an updated version of its R1 reasoning model, dubbed R1-0528, on the developer platform Hugging Face. Although the company has yet to issue an official announcement or provide technical specifications, the update appears to mark another move by DeepSeek to intensify competition with leading U.S. players such as OpenAI.
Despite the lack of public details, third-party benchmarks indicate the new R1 model performs impressively. According to the LiveCodeBench leaderboard—developed by researchers from UC Berkeley, MIT, and Cornell—R1-0528 ranks just behind OpenAI’s o4 mini and o3 models in code generation capabilities. It also outperforms competitors like xAI’s Grok 3 mini and Alibaba’s Qwen 3, reinforcing DeepSeek’s growing reputation in the global AI landscape.
The update was first reported by Bloomberg, citing a DeepSeek representative who informed users in a WeChat group that a “minor trial upgrade” had been completed and was ready for testing. This low-key rollout comes amid mounting anticipation for the release of DeepSeek’s R2 model, which was originally slated for May, as well as recent updates to its V3 large language model earlier this year.
DeepSeek has been a major disruptor since its R1 model debuted in January, challenging the notion that China lags behind the U.S. in AI due to export restrictions and limited access to advanced chips. The release sent shockwaves through global tech markets and prompted Chinese tech giants like Alibaba and Tencent to accelerate the release of their own models in response.
As DeepSeek continues to close the performance gap with U.S. AI leaders—often at a fraction of the cost, it is influencing market strategies around the world. U.S. firms like Google and OpenAI have responded by cutting prices and launching more resource-efficient models, highlighting how DeepSeek’s rapid progress is reshaping the competitive landscape of global AI development.
Source: Reuters