DeepSeek said its R1 beats or is on par with OpenAI’s o1 in several coding, math, and reasoning benchmarks. Credit: Mojahid Mottakin/Shutterstock.

China-based AI startup DeepSeek has released an open-source version of its reasoning model, DeepSeek-R1, claiming it matches OpenAI’s o1 on certain AI benchmarks.

Said to challenge OpenAI’s dominance in the AI sector with its competitive performance and cost-effective pricing, R1 is the firm’s latest open-source reasoning model, following a preview launch in November 2024, reported ZDNET.

Go deeper with GlobalData

Data Insights

The gold standard of business intelligence.

Find out more

Related Company Profiles

View all

DeepSeek’s R1 is made available on the Hugging Face platform under the MIT license, allowing unrestricted commercial use, reported TechCrunch.

In a post on social media X, the company said that R1 beats or is on par with OpenAI’s o1 in benchmarks such as AIME, MATH-500, and SWE-bench Verified.

In addition to the 671-billion-parameter model, DeepSeek has introduced six smaller “distilled” versions, with the smallest comprising 1.5 billion parameters, enabling operation on local devices.

Recently, researchers from University of California in Berkeley also introduced an open-source model that rivals o1-preview, achieving this milestone in just 19 hours and at a cost of approximately $450.

How well do you really know your competitors?

Access the most comprehensive Company Profiles on the market, powered by GlobalData. Save hours of research. Gain competitive edge.

Company Profile – free sample

Thank you!

Your download email will arrive shortly

Not ready to buy yet? Download a free sample

We are confident about the unique quality of our Company Profiles. However, we want you to make the most beneficial decision for your business, so we offer a free sample that you can download by submitting the below form

By GlobalData
Visit our Privacy Policy for more information about our services, how we may use, process and share your personal data, including information of your rights in respect of your personal data and how you can unsubscribe from future marketing communications. Our services are intended for corporate subscribers and you warrant that the email address submitted is your corporate email address.

R1’s pricing is said to be disruptive, offering API access at $0.14 per million tokens, significantly undercutting OpenAI’s $7.50 rate for the same amount of data.

OpenAI currently promotes ChatGPT Pro with unlimited o1 access for $2,400 annually.

R1, however, has limitation as it must comply with China’s internet regulations. This require R1’s responses to align with “core socialist values.”, reported  TechCrunch.

In a separate announcement, DeepSeek has developed a free, open-source large-language model using reduced-capability chips, H800s, from NVIDIA, reported CNBC.

The model has been developed in two months and at a cost of less than $6m.

At the World Economic Forum in Davos, Switzerland, Microsoft CEO Satya Nadella was quoted by CNBC as saying: “To see the DeepSeek new model, it’s super impressive in terms of both how they have really effectively done an open-source model that does this inference-time compute, and is super-compute efficient.

“We should take the developments [coming] out of China very, very seriously.”