Skip to main content

Tencent Cloud Slashes AI Model Prices by Up to 97.5% in Major Shakeup

Tencent Cloud Drops AI Model Prices in Bold Market Move

In a move that could reshape China's AI development landscape, Tencent Cloud announced sweeping price cuts for its DeepSeek-V4 series of large language models. Effective June 3, developers will see costs drop by as much as 97.5% for certain services - a reduction that makes these powerful AI tools dramatically more accessible.

Image

Deep Discounts Across the Board

The flagship DeepSeek-V4-Pro model leads the price-cutting charge with a 75% reduction in both input and output inference costs. But the real shocker comes in cache hit pricing, which plunges a staggering 97.5% - making repeated queries vastly more economical for high-volume users.

"This isn't just incremental change - it's a complete rethinking of AI cost structures," noted industry analyst Li Wei. "For startups and researchers operating on tight budgets, these cuts could be transformative."

The lightweight DeepSeek-V4-Flash model also gets a major 90% price slash on cache hits, positioning it as an affordable option for latency-sensitive applications like chatbots and real-time analytics.

Cutting-Edge Tech Meets Competitive Pricing

Behind the pricing headlines lies impressive technical capability. The DeepSeek-V4 series uses an advanced mixture-of-experts (MoE) architecture, packing 1.6 trillion parameters while handling context windows up to 1 million tokens - enough to process entire books in a single query.

Tencent's pricing overhaul coincides with its shift from free beta testing to full commercialization. By matching DeepSeek's permanent price reductions, the company clearly aims to capture market share as enterprises begin deploying AI at scale.

"We're seeing a pricing war erupt in China's cloud AI sector," observed tech journalist Zhang Ying. "With Baidu and Alibaba also slashing costs, this could accelerate AI adoption across industries from finance to manufacturing."

Key Points:

  • 97.5% price cut on DeepSeek-V4-Pro cache hits
  • 75% reduction in standard inference costs
  • 90% discount for DeepSeek-V4-Flash cache usage
  • 1.6 trillion parameter models now more accessible
  • Pricing matches DeepSeek's permanent reductions
  • Move intensifies China's cloud computing competition