As artificial intelligence (AI) continues to permeate banking operations, token consumption has emerged as a pivotal metric for gauging the scale and activity of AI applications, now prominently featured in the semi-annual reports of listed banks. Data released by leading financial institutions such as the Industrial and Commercial Bank of China, China Merchants Bank, and Ping An Bank reveals a significant transition: AI applications in the banking sector have evolved from initial technological exploration to comprehensive implementation. This expansion into diverse application scenarios has triggered a surge in token consumption, marking a shift of large-scale AI models from the 'pilot' phase to 'high-frequency production use'.