Create Chat Prefix Completion (Beta)
Query
Comma-separated response fields to include.
Body
Model ID to use
Conversation messages; last message must be assistant with prefix=true
Sampling temperature (0-2)
Maximum tokens to generate
Nucleus sampling (0-1)
Stop sequences
Enable streaming response
Include usage chunk when stream=true
Whether to return log probabilities
Response format type
Penalize repeated tokens (-2 to 2)
Penalize tokens already in text (-2 to 2)
Reasoning mode type
JSON array of tool/function definitions for function calling
Tool choice mode or JSON object
Number of most likely tokens to return (requires logprobs=true)
Role of each message
Content of each message
Optional participant name
Tool call ID for tool-role messages
Enable prefix completion on assistant message
Tool type
Function definition object
Function name
Function description
JSON Schema describing function parameters
Enable strict schema adherence
Response format object