The open-source large language model field has once again received significant news. The DeepSeek platform recently quietly released the latest 0528 version of the R1 model, which has attracted widespread attention in the AI developer community.
Notably, DeepSeek adopted a rather low-key release strategy this time. There was no detailed technical documentation released, nor any form of launch event held. Instead, they chose to "quietly" make the model available for download. As per usual practice, the official team is expected to release a full model card soon, which will disclose more information about the new version’s features and technical details.
Image source note: Image generated by AI, licensed through Midjourney
Despite limited official information, the reaction from the developer community has been extremely enthusiastic. Many technical enthusiasts have begun testing and evaluating this latest version. Initial test results on the industry-recognized code testing platform Live CodeBench show that the 0528 version of the DeepSeek R1 model exhibits remarkable performance, with overall capabilities already comparable to OpenAI's newly released o3 high-end model version.
Besides performance, the community users also conducted in-depth testing on the interaction style of the new R1 version. Test results indicate that this model demonstrates high similarity with the OpenAI o3 model in terms of dialogue mode, reasoning logic, and overall user experience, further confirming DeepSeek's rapid progress in large language model technology.
This release event once again highlights the intense competitive landscape in the open-source large language model field, while also showcasing the strong capabilities of Chinese AI teams in top-tier model development. With more test results being published and official detailed information released, the 0528 version of the DeepSeek R1 model is expected to hold an important position within the open-source AI ecosystem.
