DeepSeek built much of its reputation by offering powerful AI models at significantly lower prices than many Western rivals. That advantage is about to shrink as the Chinese AI company prepares a major increase in API pricing for its V4 models. The company has i ntroduced a new peak and off-peak pricing system alongside the release of DeepSeek V4 Pro. DeepSeek says the change is intended to allocate computing resources more efficiently by encouraging customers to move workloads away from periods of high demand. The new rates take effect tomorrow on August 16.

The company has i ntroduced a new peak and off-peak pricing system alongside the release of DeepSeek V4 Pro. DeepSeek says the change is intended to allocate computing resources more efficiently by encouraging customers to move workloads away from periods of high demand. The new rates take effect tomorrow on August 16.

DeepSeek V4 Pro currently costs $0.87 per million output tokens. Under the new system, that figure will climb to $3.96 per million output tokens during peak hours, an increase of more than four times.

Customers running workloads during off-peak periods will pay half the peak rate at $1.98 per million output tokens. DeepSeek defines peak periods as 06:00–09:00 PKT and 11:00–15:00 PKT.

The cheaper DeepSeek V4 Flash is also getting a substantial increase. Its current output price of $0.28 per million tokens will rise to $1.32 during peak hours and $0.66 during off-peak hours.

The current low V4 Pro pricing originally came from a 75% promotional discount that was scheduled to end on May 31.

DeepSeek surprised customers in May by announcing that it would make the discounted rate permanent, reducing V4 Pro pricing to one-quarter of its original level. The upcoming August increase effectively reverses that decision less than three months later.

Despite the increase, DeepSeek V4 Pro remains considerably cheaper on output tokens than several competing flagship AI models.

Moonshot AI’s Kimi K3 costs around $15 per million output tokens, while OpenAI’s flagship GPT-5.6 Sol costs $30 per million output tokens at standard API pricing.

However, DeepSeek no longer wins every price comparison. OpenAI’s lower-cost GPT-5.6 Luna costs $1.20 per million output tokens, putting it slightly below DeepSeek V4 Flash’s new $1.32 peak-hour rate.

Get the latest tech news, telecom insights, and product launches wherever you prefer.

Add ProPakistani to Preferred Sources and see more of our stories in Google Search and Top Stories.

Technology and Automotive Specialist covering the latest cars, smartphones, AI breakthroughs, and...