(Yicai) Feb. 13 -- Industry insiders have deemed Chinese small and medium lenders' decision to integrate artificial intelligence startup DeepSeek's large language model as a cost-effective AI solution to narrow the technological gap between them and larger banks.

DeepSeek offers a cost-effective intelligent business solution to small and medium banks, enabling them to quickly deploy and apply LLMs without significant investments in technological research and development, said Zeng Gang, director and chief expert at the Shanghai Institution for Finance and Development.

DeepSeek introduced its LLM DeepSeek-R1 on Jan. 20. The Hangzhou-based company offers application programming interface services priced at CNY1 (14 US cents) per million tokens of input and CNY16 (USD2.20) per million tokens of output, significantly lower than OpenAI's latest LLM O1's rates of USD15 and USD60, respectively.

Given the substantial resources required for self-developed LLMs, small and medium banks may see their technological gap with larger banks widen. Instead, integrating an already developed and low-cost LLM like DeepSeek-R1 gives them an opportunity to narrow the gap.

Bank of Beijing, Bank of Chongqing, Bank of Jiangsu, Chongqing Rural Commercial Bank, Postal Savings Bank of China, and Suzhou Merchants Bank are among the Chinese banks that have recently announced the successful deployment of DeepSeek-R1. Except for PSBC, all others are medium-sized lenders.

With DeepSeek-R1, application scenarios such as intelligent risk control, contract verification, and customer insights can be rapidly launched, greatly shortening the process from technology introduction to practical use, Zeng noted.

"DeepSeek's LLM technology boasts strong interface capabilities, efficient computational performance, and relatively low interface costs, making it highly suitable for high-frequency calls and practical applications in specific scenarios," an insider at a city commercial bank told Yicai. "This would provide strong support for the application and development of AI in small- and medium-sized lenders."

Suzhou Merchants Bank's intelligent decision-making system, which was built using DeepSeek's LLM series, has been applied in over 20 business scenarios, including credit risk control and anti-fraud monitoring, with the efficiency of generating due diligence reports improved by 40 percent, an insider told Yicai.

However, the application of DeepSeek-R1 in banks is still in the exploratory phase. The LLM's current applications are mainly focused on internal work scenarios, such as intelligent code writing, internal AI office work, and middle-office operations management, Yicai learned.

Even though DeepSeek-R1 has improved employee efficiency, it has not yet been applied to core business areas such as account transactions. Due to the extremely high requirements for data accuracy and to ensure interpretability and strict risk control, financial institutions tend to be cautious when applying LLMs, a financial industry insider told Yicai.

Editor: Futura Costaglione