DeepSeek offers 75% discount on AI new model
27 Apr 2026

DeepSeek, a leading AI lab based in Hangzhou, recently launched its latest flagship model, the DeepSeek-V4-Pro.

The company is offering a massive 75% discount on the new model to developers.

The move comes as part of an aggressive pricing strategy aimed at driving adoption and competing with other players in China's rapidly evolving artificial intelligence industry.

Input cache hits fees slashed
Fee reduction

DeepSeek dominates the market as the low-cost leader with rates between $0.28 and $0.5 per million tokens, while OpenAI's GPT-5 class sits in the mid-to-high range at $2.50 to $10, and Anthropic's Claude models command the highest premium at $3 to $25 per million tokens.

DeepSeek is also cutting down fees for input cache hits across its range of AI platforms. The move will bring down costs for regular users making similar or repeated requests by up to 90%.

Pricing strategy amid global competition
Market competition

DeepSeek's aggressive pricing strategy comes at a time when global tech giants like OpenAI, Anthropic, and Google's parent company Alphabet are launching new products at a premium.

Chinese AI companies are now offering discounts to lure users and speed up adoption in this competitive global market.

DeepSeek hopes that its pricing, accessibility, and advanced features will set its models apart for next-gen developers and enterprise users.

Model's features and user benefits
Model capabilities

The DeepSeek-V4 model comes with a large context window, allowing it to process complex code bases and long documents.

It also supports easy integration of Claude Code, OpenClaw, and OpenCode. This feature makes it easier for users to work within the broader AI ecosystem.

Akshar Keremane, co-founder of Bangalore-based AI start-up O-Health, praised these features for lowering barriers for developers, start-ups, and small enterprises.