xAI has announced that its Chat Completions endpoint is now a legacy feature. New capabilities will be delivered through the Responses API, which xAI recommends for interacting with its models. Developers currently using Chat Completions are advised to migrate to the Responses API.
Key Points
- The Chat Completions endpoint is a legacy feature, with new capabilities delivered via the Responses API.
- The Responses API is the recommended method for interacting with xAI models.
- The response format differs between the two APIs; Chat Completions returns content in
choices[0].message.content, while the Responses API uses anoutputarray with typed items. - For multi-turn conversations, Chat Completions requires resending the entire history with each request, whereas the Responses API allows continuing a conversation using
previous_response_id. - Streaming outputs are supported by all models with text output capability, including grok-4.5, and can be enabled by setting
"stream": truein the request. - Reasoning models, such as grok-4.5, can think through problems step-by-step and excel at numerical and logical tasks.
- Grok-4.5 supports a
reasoning_effortparameter to control the model's thinking process before responding.
Context
According to xAI documentation, the Chat Completions endpoint is widely used for tasks such as summarizing articles, generating creative writing, answering questions, providing customer support, and assisting with coding. However, the company is shifting its focus to the Responses API for future developments. The Responses API offers enhancements like a different response format and a more efficient way to handle multi-turn conversations by allowing developers to reference previous responses without resending full history.
Why It Matters
This change impacts developers building with xAI models, requiring them to update their API integrations to leverage new features and maintain compatibility. The Responses API offers improved conversational flow and structured outputs, which can lead to more efficient and robust applications.
What To Do
- Review the xAI documentation for the Responses API to understand its structure and capabilities.
- Compare the response formats of the Chat Completions and Responses API to plan for necessary code adjustments.
- Test the
previous_response_idfeature in the Responses API for managing multi-turn conversations. - Explore the
reasoning_effortparameter available with grok-4.5 to fine-tune model behavior for complex tasks.
Keep Exploring
/atlas/grok-family /techniques/output-schema /techniques/multishot