Instant Integration of DeepSeek-V4 via Unified API Interfaces
Seamlessly migrate to the latest models using your existing OpenAI or Anthropic SDKs.
Update your API configuration to target DeepSeek-V4 models directly through existing OpenAI or Anthropic interfaces for a frictionless upgrade.
The Scenario
You are a developer currently using the older DeepSeek-chat models and want to upgrade to the latest V4 architecture without rewriting your entire codebase.
Before & after
Developers had to wait for API providers to update their infrastructure or manually migrate to entirely new platforms with different syntax. Time cost: 1–2 hours.
Developers can switch the endpoint URL and model to `deepseek-v4-pro` to use the latest capabilities with zero architectural changes. Time cost: < 2 minutes.
The Prompt
Change your base_url to "https://api.deepseek.com" and update your model parameter to "deepseek-v4-pro". If you want to use the thinking mode, ensure you are referencing the correct mode as specified for your specific version.
The DeepSeek API has introduced V4-Pro and V4-Flash models that are accessible via both OpenAI ChatCompletions and Anthropic interfaces, simplifying integration for users of various SDKs.
Source
Change Log | DeepSeek API Docs"The DeepSeek API now supports V4-Pro and V4-Flash, available via both the OpenAI ChatCompletions interface and the Anthropic interface."