DeepSeek, known for undercutting Western rivals on AI pricing, is preparing customers for a major cost shift. With the rollout of its V4 Pro model, the company is increasing API rates fourfold and introducing peak and off-peak pricing to, in its words, "allocate resources more reasonably."

Starting August 16, using V4 Pro will cost $3.96 per million output tokens during peak hours—more than four times the current $0.87 rate. Off-peak usage drops that to $1.98. The lighter V4 Flash model jumps from $0.28 to $1.32 per million output tokens at peak, and $0.66 off-peak.

The old pricing was originally a promotion ending May 31, which DeepSeek had said it would make permanent, but it has since reversed course. Even with the hike, V4 models remain cheaper than many competitors: Moonshot's Kimi K3 runs $15 per million output tokens, and OpenAI's GPT-5.6 Sol costs $30. However, OpenAI's budget option, GPT-5.6 Luna, is $1.20—less than DeepSeek's V4 Flash at peak hours.

Source: Engadget