DeepSeek is ramping up competition in China’s AI sector by aggressively slashing prices on its newly launched flagship model, intensifying pressure on Silicon Valley rivals.

The Hangzhou-based AI lab is offering a 75 per cent discount to developers using DeepSeek-V4-Pro, which was released last week after months of anticipation.

It has also cut fees for input cache hits across its AI platforms to just a tenth of previous rates, sharply reducing costs for users who submit repeated or similar queries.

The pricing strategy risks reigniting a price war that began after DeepSeek disrupted the market with its R1 model last year.

The latest move comes as OpenAI, Anthropic, and Google accelerate new model releases, though access to their systems often remains costly for developers.

Chinese AI companies are cutting prices to attract users and accelerate adoption in an already crowded global AI market, as they seek to reshape the competition between the US and China in artificial intelligence.

DeepSeek is betting that lower pricing, greater accessibility, and advanced features will help its models stand out for next-generation developers and enterprise customers.

A key feature of DeepSeek-V4 is its large context window — the amount of information a model can process at once. This capability allows it to handle complex tasks such as analysing large codebases and long documents.

The model also supports integration with tools like Claude Code, OpenClaw, and OpenCode, making it easier for developers to plug into a wider AI ecosystem.