DeepSeek is raising prices on its flagship V4 models by up to 4x, effective August 16. V4-Flash output goes from $0.28 to $1.32/M tokens at peak hours. V4-Pro output jumps from $0.87 to $3.96. Bloomberg reports the price hike comes "ahead of a potential initial public offering." The company faces what SCMP called "challenges to maintain its aggressively low prices amid fierce competition," with AI developer Michael Guo arguing that "new American models including Meta's Muse Spark and OpenAI's GPT-5.6 Luna were competitive with DeepSeek models in both capabilities and price."

DeepSeek Raises AI Prices More Than Fourfold DeepSeek is sharply increasing prices for its flagship V4 AI models from August 16, moving its costs closer to those of major competitors. The V4-Flash model will rise to $1.32 per million output tokens during peak hours, from $0.28 previously, while V4-Pro will increase to $3.96, from $0.87. DeepSeek says the new dynamic pricing is intended to better allocate computing resources and encourage users to shift workloads to off-peak periods. Despite the increases, its services remain cheaper than some rivals, but the move could ease concerns over the impact of DeepSeek’s ultra-low pricing on the broader AI industry.

DeepSeek is sharply raising prices for its V4 models starting Aug. 16. V4-Flash output pricing during peak hours rises to $1.32 per 1M tokens from $0.28, more than 4x higher. V4-Pro will cost $3.96 per 1M output tokens at peak vs. $0.87 currently. Off-peak pricing will be half those rates, as DeepSeek shifts to dynamic pricing to move usage away from congested periods.