DeepSeek
DeepSeek V4 Pro
DeepSeek V4 Pro — 1.6T-parameter MoE model with 49B active parameters, 1M context, advanced reasoning, coding, and long-horizon agentic workflows.
deepseek-v4-proPricing
~99.9%- Input
- $0.001014 / 1,000,000 tokens
- Output
- $0.002028 / 1,000,000 tokens
- Cache read
- $0.000084 / 1,000,000 tokens
- Cache write
- $0.001014 / 1,000,000 tokens
Discount history
Model details
- Input
- text
- Output
- text
- Context window
- 1,000,000 tokens
- Maximum output
- 128,000 tokens
- Availability
- $3.9KUncapped capacity
- Updated
- 11/8/2026, 3:58:06 AM UTC
Capacity is reserved when your request is accepted.
Capabilities
ChatStreamingTool callingReasoning
Supported parameters
max_tokenstemperaturetop_pfrequency_penaltypresence_penaltyrepetition_penaltymin_plogit_biaslogprobstop_logprobsstoptoolstool_choiceresponse_formatseedstructured_outputsinclude_reasoningreasoning