Page Loader
DeepSeek updates R1 reasoning model, challenging OpenAI and Google's dominance
Updated R1 reasoning model ranked just behind OpenAI's o3 reasoning model

DeepSeek updates R1 reasoning model, challenging OpenAI and Google's dominance

May 29, 2025
11:02 am

What's the story

Chinese artificial intelligence (AI) start-up DeepSeek has launched an update to its R1 reasoning model. The new version, dubbed R1-0528, was unveiled on developer platform Hugging Face. However, no official announcement or details about the model have been released yet. This move intensifies competition with US-based companies like OpenAI and Google.

Performance ranking

DeepSeek's model ranks high on LiveCodeBench leaderboard

The updated R1 reasoning model from DeepSeek has been ranked just behind OpenAI's o4 mini and o3 reasoning models on code generation by the LiveCodeBench leaderboard. The benchmark was developed by researchers from UC Berkeley, MIT, and Cornell. It also outperformed xAI's Grok 3 mini and Alibaba's Qwen 3 in terms of performance.

Market impact

DeepSeek's R1 model challenges US AI advancements

DeepSeek's launch of the R1 model in January challenged the notion that US export controls were stifling China's AI progress. The company released AI models that rivaled or even surpassed industry-leading ones in the US, at a fraction of the cost. This move sent tech shares outside China plummeting and questioned the belief that scaling AI requires vast computing power and investment.

Industry response

Chinese tech giants respond to DeepSeek's AI models

In response to DeepSeek's groundbreaking R1 model, Chinese tech giants Alibaba and Tencent have launched their own models claiming to surpass DeepSeek's. Meanwhile, Google and OpenAI have also adjusted their strategies. Google's Gemini has introduced discounted access tiers while OpenAI has reduced prices and launched an o3 Mini model that requires less computing power.

Upcoming releases

DeepSeek's future plans and developments

DeepSeek is now expected to launch R2, a successor to the R1 model. Reports in March indicated that R2's release was initially scheduled for May. The company also upgraded its V3 large language model in March. These developments show DeepSeek's commitment to pushing the boundaries of AI technology amid stiff global competition.