DeepSeek
deepseek-v4-flash
DeepSeek V4 Flash delivers high speed performance and efficiency for modern AI tasks. It excels in reasoning, coding, and tool use with a 1M context window. The model supports both non thinking and thinking modes, providing flexibility for agentic workflows and complex implementations. It includes support for JSON output and tool calling.
🧠Reasoning🔧Tool calling⚡Caching
Pricing per 1M tokens
Input
$0.14
Output
$0.28
Cache write
$0.14
Cache read
$0.03
Specifications
Context window1M tokens
Max output384K tokens
API typechat
AddedApr 24, 2026
Model IDdeepseek/deepseek-v4-flash
Privacy & data
Data retentionYes
Used for trainingUnknown
Provider location🇨🇳 China
Privacy policyDeepSeek Privacy Policy →