DeepSeek reduced the cost of accessing the DeepSeek-V4-Pro model by 75 %

DeepSeek cuts prices and expands capabilities for developers

The Chinese company DeepSeek announced two significant changes that could reshape the AI market landscape.

1. 75 % discount on the new model

The flagship AI model *DeepSeek‑V4‑Pro* is offered at a three‑quarters price reduction. This makes it accessible to a broader audience, including startups and small businesses.

2. Lowering repeat query costs

By caching input data, DeepSeek has reduced the cost of subsequent and similar queries across all its platforms by tenfold. This approach cuts users’ operating expenses.

Why it matters
- Resumption of price wars

After DeepSeek drew Silicon Valley’s attention with its R1 model early last year, competitors—OpenAI, Anthropic, and Google—began releasing new products. However, their prices remain high. The cost cut from DeepSeek could spark a new pricing race.

- Attracting developers

Chinese companies expect lower rates to accelerate developers’ migration to their platforms and shift the balance of power in tech rivalry with the United States.

More than just price
DeepSeek is not limited to discounts. The *V4* model has:

- A large context window – up to one million tokens, enabling processing of complex codebases and large documents.

- Integration with other AI ecosystems – it easily connects to Claude Code, OpenClaw, and OpenCode, simplifying interaction with various services.

> “Pricing, open source, and a one‑million‑token context window lower the entry barrier for developers, startups, and small businesses,” said Akshar Keremane, co‑founder of O‑Health.

> According to him, such models were previously inaccessible to most experiments.

Thus, DeepSeek combines attractive pricing with advanced technical features, which could become a decisive factor when choosing an AI development platform.