Media365
The paper suggested that DeepSeek retained all 18 scientists behind its AI model development efforts, as well as many of the R1 project’s 176 contributors, despite fierce competition for talent in China’s AI industry.
The Hangzhou-based start-up first published the technical paper for R1 on January 22 last year, days after its new model made waves globally for achieving performance that was close to those of its US peers, while being trained at a fraction of the cost spent by those larger firms.
The updated paper added 64 pages to the original technical document. Much of the new content about R1’s development was featured in DeepSeek’s landmark peer-reviewed article, published in the British journal Nature in September.
The author lists of the three papers – the first version, the Nature article and the latest version – showed insights into the changes of DeepSeek’s talented research team, which has been one of the most closely watched in China’s AI sector.
Based on the asterisks next to the names of contributors who were no longer affiliated with DeepSeek, the number of asterisks had changed from six to 10 to five over the course of a year. That covered January and September last year and January 2026.
It suggested that DeepSeek managed to retain much of its AI talent a year after the company achieved international fame.
