DeepSeek’s official API documentation lists deepseek-v4-pro at $0.435 per million cache-miss input tokens and $0.87 per million output tokens, with a 1 million-token context length. A Price Per Token catalog page updated June 3 mirrors those figures and cites benchmark sources including Artificial Analysis and Huggi...
DeepSeek has listed official API pricing for deepseek-v4-pro, setting cache-miss input at $0.435 per million tokens and output at $0.87 per million tokens, according to the company’s API documentation.
DeepSeek’s “Models & Pricing” documentation lists deepseek-v4-pro with three relevant token prices: $0.003625 per million cache-hit input tokens, $0.435 per million cache-miss input tokens, and $0.87 per million output tokens. The same DeepSeek API documentation states that the model supports a 1 million-token context length.
A Price Per Token catalog page titled “DeepSeek V4 Pro API Pricing 2026” reports the same headline rates of $0.435 per million input tokens and $0.870 per million output tokens. Price Per Token says the catalog page was updated on June 3 and also lists a 1.0 million-token context window.
The distinction between cache-hit and cache-miss input pricing matters for API users comparing costs. DeepSeek’s documentation prices cache-hit input tokens far lower than cache-miss tokens, at $0.003625 per million versus $0.435 per million. That means the effective cost for repeated prompts or reused context may differ substantially from one-off requests, depending on how the API applies caching in a given workload.
DeepSeek’s separate “DeepSeek V4 Preview Release” notice says the company announced the V4 preview release with DeepSeek-V4-Pro included. In that release note, DeepSeek describes DeepSeek-V4-Pro as having 1.6 trillion total parameters and 49 billion active parameters. The same DeepSeek announcement says the model has a 1 million-token context length, open weights, and same-day API availability.
Those details position V4 Pro as a large mixture-style model in DeepSeek’s public lineup, but the provided source material does not include independent testing results beyond references to benchmark sources. Price Per Token says its catalog page includes benchmark data from Artificial Analysis and Hugging Face’s Open LLM Leaderboard, but the excerpt does not provide specific scores or rankings.
For developers evaluating model economics, the clearest source-backed comparison point is the official per-token pricing published by DeepSeek. At the listed rates, a million cache-miss input tokens cost $0.435 and a million output tokens cost $0.87. Cache-hit input tokens are priced at $0.003625 per million tokens in DeepSeek’s documentation.
The 1 million-token context length is also a key operational detail for applications that need to process long documents, extended chat histories, codebases, or retrieval-augmented prompts. DeepSeek’s pricing page and preview release both state that deepseek-v4-pro supports that context length.
Price Per Token’s catalog entry provides a secondary reference that aggregates the model’s pricing and context-window information and notes benchmark sources. However, the official DeepSeek API documentation remains the primary source for the API pricing figures cited here.
The sources provided do not specify regional availability, rate limits, service-level guarantees, or whether preview-release conditions differ from general availability. They also do not provide enough benchmark detail to compare DeepSeek-V4-Pro’s performance against other frontier or open-weight models in this article.
For now, the documented facts are narrow but useful: DeepSeek lists deepseek-v4-pro at $0.003625 per million cache-hit input tokens, $0.435 per million cache-miss input tokens, and $0.87 per million output tokens, with a 1 million-token context length; DeepSeek’s V4 preview note describes the model as open weight with 1.6 trillion total parameters and 49 billion active parameters.
DeepSeek has listed official API pricing for deepseek v4 pro, setting cache miss input at $0.435 per million tokens and output at $0.87 per million tokens, according to the company’s API documentation.
The same DeepSeek API documentation states that the model supports a 1 million token context length.
A Price Per Token catalog page titled “DeepSeek V4 Pro API Pricing 2026” reports the same headline rates of $0.435 per million input tokens and $0.870 per million output tokens.
Continue reading
