DeepSeek Launches V4.1 Flash Model, Redirects V4 Pro Requests

TechNode company DeepSeek formally launched its V4.1 Flash AI model on September 10, introducing a new pricing structure for its Flash series. The company announced that V4.1 Flash has outperformed the previous V4 Pro model in internal and external testing across performance, cost, speed, and total completion time.

Effective September 14 at noon Beijing time, all requests previously directed to V4 Pro will be automatically routed to V4.1 Flash. These requests will subsequently be billed according to the Flash-series rates. This strategic shift aims to provide users with enhanced efficiency and cost-effectiveness.

The new pricing model includes off-peak rates of RMB 0.02 per million tokens for cache-hit inputs, RMB 1 for cache-miss inputs, and RMB 4 for outputs. During peak hours, these prices will be doubled. This tiered pricing is designed to manage demand and offer competitive rates.

This transition signals DeepSeek's commitment to optimizing its AI service offerings, ensuring users benefit from the latest advancements in processing speed and resource utilization. The company's move reflects an ongoing trend in the AI sector towards more specialized and efficiently priced models.