Kimi API Announces Major Compatibility Enhancement
Developers working with AI-powered coding assistants often face integration hurdles due to incompatible API formats across different services. This typically requires building custom adapters or proxy layers, adding unnecessary complexity to projects.
Native Support for Leading API Standards
A recent update addresses this issue head-on. Kimi API now offers native support for two major industry-standard formats:
- OpenAI Responses API Format: Developers can now call Kimi's models within Codex environments using the exact same request structure as native OpenAI models.
- Anthropic Messages API Format: Similarly, integrating Kimi models into Claude Code workflows now aligns seamlessly with the standard Claude API call format.
The key benefit here is elimination of friction. There's no need to craft unique request bodies for Kimi or maintain a local proxy server for format translation.
Direct Model Access Simplified
To get started, developers simply configure their custom model provider settings in Codex or Claude Code to point to the Kimi API endpoint and specify their chosen model. Available models now include:
- kimi-k3
- kimi-k2.7-code-highspeed
- kimi-k2.7-code
- kimi-k2.6
Imagine you have an existing Codex workflow. You can now switch your model configuration to "kimi-k2.7-code-highspeed" without rewriting your core application logic.
This update dramatically lowers the barrier to entry for testing and integration. Evaluating different models for code generation tasks or incorporating new AI capabilities into ongoing projects becomes a straightforward process. Teams employing multi-model strategies will find management and switching between providers significantly more efficient.