DeepSeek slashes AI model prices 97% below OpenAI's GPT-5.5

DeepSeek has significantly slashed the prices of its AI models, including the newly launched V4, which is now priced at 97 percent less than similar offerings from OpenAI. On Sunday, the Chinese firm revealed a reduction in costs for 'input cache hits' for API users to just one-tenth of the previous rate, resulting in a minimum input expense of approximately US$0.14 per million tokens. These price reductions are effective immediately and will remain in place indefinitely. The DeepSeek-V4-Pro is now available for as little as US$0.0036 per million input tokens, compared to OpenAI's GPT-5.5, which charges US$0.5 per million cached input tokens. Since input typically exceeds output by three times, the cost per conversation on GPT-5.5 is 32 times higher than that of DeepSeek-V4. This strategy aims to draw in more enterprise clients, developers, and agent-based users, potentially igniting a price competition in the AI sector.