DeepSeek Just Raised Its AI Prices Up to 14x

DeepSeek built its global reputation on being the cheap alternative to Western AI labs. That reputation just took a sharp turn. The Chinese AI company formally released its V4 Pro model on August 13, 2026, and priced it dramatically higher than its own budget-tier model, a genuine reversal for a company whose entire pitch had been affordability.

Cr: itdaily.com

How Much Prices Actually Went Up

DeepSeek-V4-Pro-0813 is priced at $1.32 per million input tokens and $3.96 per million output tokens, according to independent benchmarking firm Artificial Analysis. That's roughly nine times the input price and fourteen times the output price of DeepSeek's own V4 Flash model, which costs $0.14 and $0.28 respectively. Output pricing alone rose more than four times its previous flat rate, and some rates increased by as much as 1,100% depending on the pricing tier.

DeepSeek also introduced peak and off-peak billing, a first for the company. During peak hours, prices sit at the higher rate; off-peak pricing runs at roughly half.

Why the Price Actually Went Up

DeepSeek says V4 Pro delivers substantially stronger performance for AI agents, tasks where the model uses tools, runs code, and completes multi-step workflows without constant human input. Benchmark results the company released show real gains: an Intelligence Index score of 53, compared to 40 for its own Flash model, along with strong results on agent-focused tests like Terminal Bench and NL2Repo.

The Bigger Business Story

The price hike lands right as DeepSeek is reportedly pursuing a new funding round that would value the company at around $74 billion, following a roughly $7.4 billion outside financing round in June, its first ever. That's a significant shift for a company that had avoided outside capital entirely until this year.

Running frontier AI models is expensive, and unlike a company such as Alibaba, which can afford to price its models near cost because it profits from the cloud infrastructure underneath, DeepSeek has no cloud business to subsidize its token pricing. Serving reasoning-heavy AI workloads means real GPU compute and real power costs, and DeepSeek's new price list appears to reflect that reality more honestly than its earlier, famously cheap pricing did.

Frequently Asked Questions

How much did DeepSeek raise its prices?

DeepSeek's new V4 Pro model costs roughly 9 times more for input tokens and 14 times more for output tokens compared to its own budget-tier V4 Flash model.

Why did DeepSeek raise its prices?

DeepSeek says V4 Pro offers substantially improved performance for AI agent tasks, and the price increase reflects the real computing cost of running a more capable, reasoning-heavy model at scale.

Is DeepSeek still cheaper than Western AI companies?

Even after the increase, DeepSeek's prices remain below many Western mid-tier models, though the gap has narrowed significantly compared to its earlier ultra-low pricing.

What is DeepSeek's current valuation?

DeepSeek is reportedly in talks for a funding round that would value the company at around $74 billion, following a $7.4 billion outside financing round in June 2026.

Does the price increase affect the free DeepSeek app?

The reported price changes apply to DeepSeek's API pricing. Free access through DeepSeek's consumer app and web interface remains available separately from API token costs.

Share :