AI Development
DeepSeek confirms official V4-Flash-0731 pricing and API features
Editorial Analysis
DeepSeek's official API pricing documentation lists DeepSeek-V4-Flash-0731 at 0.14 dollars per million input tokens on cache miss and 0.28 dollars per million output tokens, with 1 million context length, up to 384K max output, support for tool calls, JSON mode, Responses API, and both thinking and non-thinking modes. DeepSeek-V4-Pro is priced higher at 0.435 dollars input cache-miss and 0.87 dollars output. The page also notes an upcoming peak/off-peak pricing policy that will double rates during specified Beijing hours, with the effective date to be announced.
At a Glance
Date
July 31, 2026
Importance
Medium
3/5
Category
inference
Axis of Change
Cost Reduction
Organizations
Models Affected
Sources