DeepSeek's flagship million-context text model for difficult reasoning, coding, long-running agents, tools, and very large outputs.
Peak: $1.32 input · $3.96 output / 1M tokens
Read guideAll models
AI model company
DeepSeek AI models
DeepSeek develops open-weight and hosted language models for reasoning, coding, agents, and long-context work. The V4 family is available through DeepSeek's applications and OpenAI- or Anthropic-compatible APIs.
About DeepSeek
Models from DeepSeek
DeepSeek's lower-cost V4 model for high-volume reasoning, coding, agents, and million-token text workloads.
Peak: $0.44 input · $1.32 output / 1M tokens
Read guide