OpenAI has claimed that DeepSeek may have distilled its AI models to create the R1 model, as per the report.
As per the report of Financial Times, OpenAI has proof that some users were using OpenAI’s application programming interface (API) to copy its models and help its competitor, which is suspected to be DeepSeek.
The company has launched its open-source DeepSeek-R1 AI model on GitHub and Hugging Face. It has gained a lot of popularity and suppressed OpenAI’s o1 models in several tests.
After investigating with its cloud partner Microsoft, OpenAI blocked those accounts.
OpenAI said in a statement to the Financial Times, “We know [China]-based companies — and others — are constantly trying to distill the models of leading US AI companies.”
The company also mentioned that it is working with the US government to protect its advanced AI models from competitors and potential threats.
AI model distillation is a technique used to transfer knowledge from a large AI model to a smaller, more efficient one. The aim is to make the smaller model perform as well as or even better than the larger one while using less computing power.
Recently, DeepSeek gained notoriety for developing its R1 reasoning model for training large AI models at a significantly lower cost than is customary.
DeeSeek, a Chinese company, developed its R1 reasoning model for training large AI models at a lower cost than usual. The company claims to have trained its model with 671 billion parameters for only $5.6 million, using 2,048 Nvidia H800 graphics cards.
