tutorial

Simplify Model Migration Using DeepSeek V4’s Dual Interface Support

Q: What prompt should I use?

# No specific prompt required; this is an implementation tip for developers. # Usage: client.chat.completions.create(model="deepseek-v4-pro", messages=[{"role": "user", "content": "[YOUR_PROMPT]"}])

Published 06:03 MYT · Jul 19

Swap to V4 models using existing OpenAI or Anthropic API libraries.

Leverage the new V4-Pro and V4-Flash models by simply changing the model ID within your existing OpenAI or Anthropic compatible codebase.

DeepSeek

The Scenario

You are migrating an existing application from OpenAI or Anthropic to DeepSeek to save costs or improve performance, and you want to avoid a total code rewrite.

Before & after

The old way

Writing custom API wrappers for every new model provider’s unique format typically takes 1-2 hours of development and testing time.

With AI

Implement the OpenAI or Anthropic compatible interface to call `deepseek-v4-flash` or `deepseek-v4-pro`. This integration takes about 10 minutes if you already use these standard libraries.

The Prompt

# No specific prompt required; this is an implementation tip for developers. 
# Usage: client.chat.completions.create(model="deepseek-v4-pro", messages=[{"role": "user", "content": "[YOUR_PROMPT]"}])

The latest V4 models support both major AI interface standards, making it easier to swap DeepSeek into existing applications.

Source

Change Log | DeepSeek API Docs

"available via both the OpenAI ChatCompletions interface and the Anthropic interface."