DeepSeek, a Chinese artificial intelligence (AI) company, plans to make parts of its online services’ code publicly available during an “open source week” event this week, according to a post on its X (formerly Twitter) account on Friday.
This move could benefit end users by providing more advanced, customizable, and affordable AI tools and services. At the same time, it may put pressure on other AI tools like ChatGPT, OpenAI, and Google Gemini to follow suit.
DeepSeek will release five code repositories (repos), which include source code and documentation that have been tested and used in their systems.
While the company has’nt details which repos it will release, its DeepSeek Open Infra affirmed that the coming releases will cover “code that moved our tiny moonshot forward” and share “our small-but-sincere progress with full transparency.” The page also pointed out to a 2024 paper detailing DeepSeek’s training related architecture and software.
The company believes open-sourcing its code will speed up progress through collaboration with the wider developer community, stating, “Every line shared builds momentum.”
DeepSeek is known for making its AI models available with flexible licenses, allowing it to challenge companies like OpenAI, which its CEO, Sam Altman, has admitted that DeepSeek has closed the technological gap, and hinted that OpenAI may open-source more of its own technology in response.
DeepSeek’s decision to open-source parts of its code aims to drive progress through collaboration, and has pressured companies like OpenAI to consider doing the same.
In January, DeepSeek made waves by releasing its low-cost R1 chatbot model, which was seen as a “Sputnik moment” in the AI industry, highlighting the potential for affordable AI solutions. This move led to a drop in shares for major AI companies.
Founded in May 2023 by Liang Wenfeng (who also founded the hedge fund High-Flyer), DeepSeek specializes in developing open-source large language models (LLMs) and has quickly gained attention for its AI advancements.
For end users, open-sourcing the code means faster innovation and new features, as developers and the broader AI community can improve and customize the software. This could result in more flexible and tailored solutions for businesses. Additionally, open-source code often lowers licensing fees, meaning users could pay less for services powered by DeepSeek’s technology.
Open-sourcing also provides more transparency, allowing users to understand how the services work and trust that any issues will be fixed quickly by the community. As more companies adopt open-source models, it could increase competition, leading to better services and lower prices for end users over time.