Rednote Joins Chinese Company Wave That Releases Open-Source AI Model
JAKARTA Rednote, one of China's most popular social media platforms also known as Xiaohongshu, has released an open-source large language language model (LLM). This move makes Rednote part of a wave of Chinese technology companies that makes their artificial intelligence model freely available.
This approach is different from many technology giants from the United States such as OpenAI and Google, which still maintain their most advanced AI model on a proprietary basis. Even so, several American companies such as Meta have also started releasing open-source models.
By opening access to its AI model, Chinese companies can demonstrate their technological capabilities, build a development community, and expand influence globally, especially in the midst of the United States' efforts to limit China's technological advances through advanced semiconductor export restrictions.
The Rednote model called dots.llm1 can now be downloaded through the Humging Face development platform. A technical paper that explains this model was also uploaded on Friday.
In the programming task, this model shows a performance comparable to Alibaba's Qwen 2.5 series, although it is still lagging behind more advanced models such as DeepSek-V3, according to the technical paper.
SEE ALSO:
Rednote, which operates similar to Instagram, allows users to share photos, videos, posts and live broadcasts. The platform received international attention earlier this year when a number of US users switched to using the app due to concerns about the potential ban on TikTok.
The company started investing in major language model development since 2023, shortly after OpenAI launched ChatGPT in late 2022. In recent months, Rednote has accelerated its efforts in the AI sector by launching Diandian, an AI-based search application that helps users find content on Xiaohongshu's main platform.
Apart from Rednote, several other companies also adopted an open-source approach, including Alibaba, which in April launched Qwen 3, the latest version of the model. DeepSeek startup also shocked the global AI industry earlier this year by releasing an affordable and competitive R1 model despite being developed at a much lower cost than Western competitors.