According to data compiled by independent research firm Melius Research, the cost of generating information using DeepSeek's AI model is approximately one-tenth of that of OpenAI's ChatGPT. Melius Research analyst Ben Reitzes suggests that DeepSeek has likely mastered techniques for reducing model training costs through more efficient memory usage and learning strategies.
DeepSeek employs a computational method that activates only the most relevant parts of the model to answer questions. This approach, coupled with its open-source nature, helps in lowering licensing costs. Reitzes indicates that competitors might respond by adjusting their capital expenditures to reduce prices.
