DeepSeek delivers the first AI capable achieving high scores in math Olympiad.

(Web Desk) – DeepSeek, a Chinese business, has produced the first-ever open AI model capable of winning gold at the famous International Mathematical Olympiad.

The Math-V2 model received a gold-level score at this year’s IMO, a feat accomplished by only 8% of human competitors, by displaying its thinking ability rather than simply producing a single solution.

DeepSeek has now made the model available to the public via the development platforms Hugging Face and GitHub, allowing anybody to run and alter it for free.

“Imagine owning the brain of one of the best mathematicians in the world for free to explore it for research, fine-tune it, optimise it, and run it on your own hardware,” said Hugging Face co-founder and CEO Clement Delangue.

“There are no constraints, no nerfing [reduction of strength], and no corporate or government to take it back. That is democratization of AI and information at its finest, literally.”

Models developed by Google DeepMind and ChatGPT maker OpenAI both received gold medals at the annual IMO, but neither firm has released them open-source.

DeepMind’s model is now available to customers of its premium Ultra package, although OpenAI has yet to publicly reveal its model.

According to the firm’s experts, DeepSeek’s model demonstrated its capacity to “self-verify,” allowing it to solve mathematical problems that had no known answers.

According to the researchers, this tackles a major barrier in present AI systems, which often advance exclusively on tasks with clearly verifiable answers.

DeepSeek debuted earlier this year, generating headlines for claiming that its large language model (LLM) was built for a fraction of the expense of competitors such as OpenAI’s ChatGPT.

It swiftly became the most popular free app in the United States, prompting President Donald Trump to label it a “wake-up call” for the internet sector.