Chinese AI startup DeepSeek has released an upgraded version of its R1 reasoning model, dubbed R1-0528, aiming to enhance logical reasoning capabilities while minimizing hallucinations (via Bloomberg).
The updated model is now available on the Hugging Face platform, intensifying competition with U.S. AI firms like OpenAI.
The R1-0528 model builds upon its predecessor by incorporating advanced training methodologies to improve reasoning tasks and reduce instances of generating inaccurate or fabricated information, commonly referred to as hallucinations.
While specific technical details remain undisclosed, benchmarking on LiveCodeBench—a collaborative project by UC Berkeley, MIT, and Cornell—indicates that R1-0528 ranks just below OpenAI’s o4 mini and o3 models in code generation tasks, outperforming competitors like xAI’s Grok 3 mini and Alibaba’s Qwen 3.
The original R1 model, launched in January 2025, was developed with significantly lower investment, leveraging innovative techniques such as reinforcement learning to automate human feedback processes. This approach has allowed DeepSeek to produce models that rival those of OpenAI and Anthropic in performance while maintaining cost efficiency.
In line with its commitment to open-source principles, DeepSeek has made the R1-0528 model accessible to developers and researchers worldwide. By providing open access to its models, DeepSeek enables a broader range of applications and facilitates advancements in various AI-driven fields.
By focusing on enhancing reasoning capabilities and reducing hallucinations, DeepSeek addresses critical challenges in AI development, particularly in ensuring the reliability and accuracy of AI-generated outputs.
Want to see more of our stories on Google?
P.S. Want to keep this site truly independent? Support us by buying us a beer, treating us to a coffee, or shopping through Amazon here. Links in this post are affiliate links, so we earn a tiny commission at no charge to you. Thanks for supporting independent Canadian media!
