DoublewordDoubleword

Model Name

deepseek-ai/DeepSeek-V4-Pro-0813

DeepSeek V4 Pro 0813

  • Type: Generation
  • Capabilities: enhanced_structured_generation, reasoning
  • Cache read: $0.13 per 1M input tokens (0.1× standard input price). See prompt caching.

Overview

DeepSeek’s official V4 Pro 0813 release, optimized for advanced reasoning, coding, tool use, and agentic workloads.

Pricing

PriorityInput Tokens (per 1M)Cache Read Tokens (per 1M)Output Tokens (per 1M)
Realtime1$1.32$0.13$3.96
Async$0.99$0.10$2.97
Batch (24h)$0.66$0.07$1.98

Playground

Open this model in the Playground.

Footnotes

  1. Realtime availability is limited. Doubleword is primarily a batch API.