AI Models
31 July 2026
DeepSeek Launches Public Beta of V4-Flash API
N
New-ZZZ desk
X @deepseek_ai · 1 week ago
DeepSeek has launched a public beta of the official V4-Flash model API. The company says that following improvements to its agentic capabilities, the model now significantly outperforms V4-Pro-Preview in performance benchmarks.
V4-Flash now natively supports the Responses API format and has been adapted for use with Codex. The model’s architecture and size remain unchanged from the preview version.
The update applies only to the DeepSeek-V4-Flash API. The V4-Pro API, as well as the models available in the app and web version of the service, remain unchanged for now.
Why it matters
- —The new API expands developers' capabilities for building AI agents powered by DeepSeek.
- —Responses API support and Codex integration make it easier to connect the model to modern development tools.
- —Improved results without changes to the model's architecture or size indicate software and API configuration enhancements.
Key facts
- The DeepSeek-V4-Flash API is now available in public beta.
- According to DeepSeek, its agentic benchmark results now surpass those of V4-Pro-Preview.
- The model natively supports the Responses API format and is adapted for Codex.
- The architecture and size of V4-Flash remain unchanged from the preview version.
- The V4-Pro API and DeepSeek models in the app and web version have not been updated.
The full text is in the original source. Here we provide a brief summary and key facts.
