DeepSeek’s new AI reasoning model R1 could substantially reduce the costs of developing AI applications. According to an analysis by Ben Dickson in VentureBeat, DeepSeek-R1 offers similar capabilities to leading models at a fraction of the price. The model costs $2.19 per million output tokens, compared to OpenAI’s o1 at $60 per million tokens. When accessed through U.S.-based providers like Together.ai and Fireworks AI, R1 still remains significantly cheaper at $8-9 per million tokens.
While the model’s capabilities are reported to be slightly below o1, they are sufficient for most enterprise applications. The release has already impacted the AI industry, leading to a selloff in major AI stocks and prompting responses from competitors. OpenAI, for instance, announced expanded access to their o3-mini model for free ChatGPT users shortly after R1’s release.
The new model has gained significant traction in the open-source community, with over 500 derivatives appearing on Hugging Face and millions of downloads. DeepSeek-R1 is available in various sizes, from 1.5 billion to 671 billion parameters, allowing organizations to run it on different hardware configurations. While questions remain about the model’s training process, its release suggests that advancing AI technology may not always require massive computing resources and datasets.