This section provides a detailed description of the parameters returned in streaming responses, using JSON Schema format.
{
"definitions": {
"ChatMessageRole": {
"enum": [
"system",
"user",
"assistant",
"tool"
],
"title": "ChatMessageRole",
"type": "string"
},
"FinishReason": {
"enum": [
"stop",
"length",
"tool_calls",
"content_filter",
"malformed_function_call"
],
"title": "FinishReason",
"type": "string"
},
"OpenAIChatCompletionStreamChoice": {
"properties": {
"index": {
"description": "The position of this choice within list of generated completions",
"title": "Index",
"type": "integer"
},
"delta": {
"$ref": "#/definitions/OpenAIDeltaMessage",
"description": "Chat completion data generated by streamed model responses"
},
"logprobs": {
"$ref": "#/definitions/OpenAIChoiceLogProbs",
"description": "Log probability information for the choice."
},
"finish_reason": {
"$ref": "#/definitions/FinishReason",
"description": "The reason generation was terminated, such as reaching a stop sequence, hitting the token limit, or invoking a tool."
}
},
"required": [
"index",
"delta"
],
"title": "OpenAIChatCompletionStreamChoice",
"type": "object"
},
"OpenAIChatCompletionTokenLogprob": {
"properties": {
"token": {
"description": "The unit Token",
"title": "Token",
"type": "string"
},
"bytes": {
"description": "List of int representing the UTF-8 bytes representation of tokens.",
"items": {
"type": "integer"
},
"title": "Bytes",
"type": "array"
},
"logprob": {
"description": "Log probability of the tokens",
"title": "Logprob",
"type": "number"
}
},
"required": [
"token",
"logprob"
],
"title": "OpenAIChatCompletionTokenLogprob",
"type": "object"
},
"OpenAIChoiceLogProbs": {
"properties": {
"content": {
"description": "List of message content tokens with log probability information.",
"items": {
"$ref": "#/definitions/OpenAIChatCompletionTokenLogprob"
},
"title": "Content",
"type": "array"
}
},
"title": "OpenAIChoiceLogProbs",
"type": "object"
},
"OpenAIDeltaMessage": {
"properties": {
"role": {
"$ref": "#/definitions/ChatMessageRole",
"description": "Role of the author of this message"
},
"content": {
"description": "Chunk message content",
"title": "Content",
"type": "string"
},
"reasoning_content": {
"description": "Reasoning content",
"title": "Reasoning Content",
"type": "string"
},
"tool_calls": {
"description": "Tool calls generated by the model",
"items": {
"$ref": "#/definitions/OpenAIDeltaToolCall"
},
"title": "Tools Called",
"type": "array"
}
},
"title": "OpenAIDeltaMessage",
"type": "object"
},
"OpenAIDeltaToolCall": {
"properties": {
"index": {
"title": "Index",
"type": "integer"
},
"id": {
"title": "Tool Id",
"type": "string"
},
"function": {
"$ref": "#/definitions/OpenAIDeltaToolCallFunction"
},
"type": {
"const": "function",
"title": "Type",
"type": "string"
}
},
"required": [
"index"
],
"title": "OpenAIDeltaToolCall",
"type": "object"
},
"OpenAIDeltaToolCallFunction": {
"properties": {
"arguments": {
"title": "Arguments",
"type": "string"
},
"name": {
"title": "Name",
"type": "string"
}
},
"title": "OpenAIDeltaToolCallFunction",
"type": "object"
},
"PromptTokensDetails": {
"properties": {
"cached_tokens": {
"description": "Total number of input tokens read from cache",
"title": "Cached Tokens",
"type": "integer"
},
"cache_write_tokens": {
"description": "Total number of input tokens used to create cache entry",
"title": "Cache Write Tokens",
"type": "integer"
}
},
"title": "PromptTokensDetails",
"type": "object"
},
"ServiceTier": {
"enum": [
"default",
"priority"
],
"title": "ServiceTier",
"type": "string"
},
"UsageInfo": {
"additionalProperties": true,
"properties": {
"prompt_tokens": {
"default": 0,
"description": "Total number of tokens in the prompt",
"title": "Prompt Tokens",
"type": "integer"
},
"total_tokens": {
"default": 0,
"description": "Total number of tokens in the completion (prompt + completion)",
"title": "Total Tokens",
"type": "integer"
},
"completion_tokens": {
"default": 0,
"description": "Total number of tokens generated in the completion",
"title": "Completion Tokens",
"type": "integer"
},
"estimated_cost": {
"description": "Total estimated cost of the completion in USD",
"title": "Estimated Cost",
"type": "number"
},
"prompt_tokens_details": {
"$ref": "#/definitions/PromptTokensDetails",
"description": "breakdown of prompt tokens usage"
}
},
"title": "UsageInfo",
"type": "object"
}
},
"required": [
"model",
"choices"
],
"title": "OpenAIChatCompletionStreamOut",
"type": "object",
"properties": {
"service_tier": {
"$ref": "#/definitions/ServiceTier",
"description": "Service tier the request was processed with",
"type": "string",
"title": "ServiceTier"
},
"id": {
"description": "Unique identifier for the completion",
"title": "Id",
"type": "string"
},
"object": {
"default": "chat.completion.chunk",
"description": "Object type, which is always chat.completion.chunk",
"title": "Object",
"type": "string"
},
"created": {
"description": "Unix system timestamp of the completion",
"title": "Created",
"type": "integer"
},
"model": {
"description": "model name",
"title": "Model",
"type": "string",
"example": "meta-llama/Llama-2-70b-chat-hf"
},
"choices": {
"description": "List of chat completion choices, can be more than one",
"items": {
"$ref": "#/definitions/OpenAIChatCompletionStreamChoice"
},
"title": "Choices",
"type": "array"
},
"usage": {
"$ref": "#/definitions/UsageInfo",
"description": "Usage info about request and response",
"type": "object",
"title": "UsageInfo"
}
}
}