Skip to main content

​
DeepSeek V4 Flash

Best price-per-token in the catalog for long-context code and JSON-mode workloads.
Model ID: deepseek-ai/DeepSeek-V4-Flash

​
About

DeepSeek V4 Flash ships an updated MoE architecture with strong code generation and JSON-mode reliability at a competitive price, with a 1M-token context window. Pricing above is provisional and will be confirmed at launch.

​
Pricing

DirectionRate (USD)
Input$0.35 / 1M tokens
Output$1.40 / 1M tokens

​
Capabilities

  • chat
  • json-mode
  • long-context

​
Specs

  • Context window: 1.0M tokens
  • License: DeepSeek License
  • Released: April 15, 2026
  • Chat and instruction following
  • Structured-output generation (JSON mode)
  • Long-document Q&A and summarization

​
Benchmarks

Benchmarks fill in during Alpha. Tracked as a follow-up to RD-562.

​
Example prompts

  • Generate a TypeScript Zod schema for this JSON sample.
  • Refactor this 600-line Python module into smaller composable units.

​
Quick start

curl https://api.corvex.ai/v1/chat/completions \
  -H "Authorization: Bearer $CORVEX_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-ai/DeepSeek-V4-Flash",
    "messages": [{"role": "user", "content": "Hello"}]
  }'
Open in the Corvex playground