The statement came in documents released alongside a peer-reviewed version of the R1 model, published today in Nature. The paper updates a preprint released in January, which describes how DeepSeek augmented a standard large language model (LLM) to tackle reasoning tasks. DeepSeek says R1 was trained mainly on Nvidia’s H800 chips, which in 2023 became forbidden from being sold to China under US export controls. The model has been “quite influential” among AI researchers, says Sun. For researchers, R1 is still very competitive, Sun says.