Race to the Bottom: DeepSeek Slashes V4-Pro API Prices by 75% After Lukewarm Market Reaction

The generative AI market is entering a phase of brutal price dumping. Following the Friday release of the `DeepSeek-V4 Preview` model, investors and markets reacted unexpectedly sluggishly. Exhausted by constant announcements, the industry is no longer buying metrics for the sake of metrics.

In response, DeepSeek made an aggressive strategic move: on April 27, 2026, the company open-sourced the model's weights and radically slashed its price list. The cost of using V4-Pro has been reduced by 75%, and the price of token caching (input cache hits) has plummeted by a factor of 10. This is a classic price war of attrition. The Chinese developer is trying to offset the technological gap (caused by the shortage of top-tier NVIDIA chips) with incredibly cheap business integration, forcing Western competitors to burn venture capital in attempts to retain corporate clients.

Source: DeepSeek / Reuters