DeepSeek has reduced prices for its V4-Flash model by up to 60%, and restored off-peak rates to the levels seen before the August price increase. The reduction applies only to the Flash tier; Pro-level pricing remains unchanged.
New lower-cost beta and native multimodality
At the same time, DeepSeek opened a beta for V4.1 Flash, positioned as a cheaper model that offers native multimodality. This move signals the company is pursuing both price and product differentiation to respond to mounting competitive pressure.
Practical impact
DeepSeek built part of its reputation on offering the cheapest frontier models. Recently, however, lower-cost competitors have multiplied. GLM 5.3 Flash has reached price parity with DeepSeek Flash and undercuts the Pro tier by a factor of three to five, while delivering nearly the same performance according to market comparisons.
That development weakens DeepSeek’s prior advantage based on low price alone. The company’s remaining lever appears to be further price reductions, coupled with the hope that a high cache-hit ratio will make discounts feel meaningful to customers.
Implications and outlook
The firm that entered the global market by undercutting competitors is now itself being undercut. With GLM 5.3 Flash and other inexpensive models on the market, competition has intensified; DeepSeek’s options are narrowing. Unless it can sustain service quality and cache efficiency, the company may need to pursue additional price cuts or product improvements to maintain its market position.
In short: DeepSeek remains committed to competing on price, but the arrival of cheaper alternatives means the company may have to keep cutting prices or otherwise change strategy to hold onto customers.
