DeepSeek has made its DeepSeek-V4-Flash-0731 model available on the Hugging Face platform, marking the official transition from a preview to a public beta version. The model's architecture and size remain unchanged, but it has been re-trained to improve its performance. The updated model is now available for developers to use and test. This upgrade is significant for developers who rely on the DeepSeek model in their applications, as it is likely to lead to more accurate and efficient results.