Chat Completions API
POST/chat/completions
Creates a model response for the given chat conversation.
Request
- application/json
Body
required
Array [
- System message
- User message
- Assistant message
- Tool message
- Latest reminder message
- Text content
- Array of content parts
Array [
- Text content part
- Image content part
- File content part
]
- Text content
- Array of content parts
Array [
- Text content part
- Image content part
- File content part
]
]
- Single stop sequence
- Stop sequence list
Array [
]
deepseek-flash: In thinking mode, the default is 1.0 and values below 0.95 are treated as 0.95. In non-thinking mode, it is fixed at 1.0 and the value you pass is ignored.deepseek-v4-pro: In thinking mode, it is fixed at 0.95 when tool calling is enabled, or 1.0 otherwise; the value you pass is ignored. Settingtool_choicetononedisables tool calling even whentoolsis provided. In non-thinking mode, the value you pass is used, with a default of 1.0.Array [
- Function tool
- DeepSeek web search tool
- DeepSeek read URL tool
]
- ChatCompletionToolChoice
- ChatCompletionNamedToolChoice
- user_id can be used to distinguish user identities on your side to help us with content safety review.
- user_id can be used for KVCache isolation for privacy management.
- user_id can be used for scheduling isolation of users on your business side.
- For more details on the user_id parameter, please refer to Rate Limit & Isolation
messages
object[]
required
Minimum items: 1
A list of messages comprising the conversation so far.
oneOf
The contents of the system message.
Possible values: system
The role of the messages author, in this case system.
An optional name for the participant. Provides the model information to differentiate between participants of the same role.
required
The contents of the user message. Either a string, or an array of content parts (for image input). See the Vision guide for details.
oneOf
string
Minimum items: 1
oneOf
Possible values: text
The type of the content part, in this case text.
The text content.
Possible values: image_url
The type of the content part, in this case image_url.
image_url
object
required
Either an http(s) URL of the image (max 8192 characters) or a base64-encoded data URL (data:image/jpeg;base64,...). Supported formats: JPEG, PNG, GIF, and WebP.
Possible values: low, high, original, auto
Controls how the image is processed. low downsamples the image to 512x512 (faster, cheaper). high, original, and auto keep the original image.
Exactly one of file_id and file_data is required: passing neither returns a 400 error ("file must have a file_id or file_data"); passing both returns a 400 error ("file cannot have both file_id and file_data").
Possible values: file
The type of the content part, in this case file.
The ID of a file uploaded via the Files API, of the form file-api-.... Mutually exclusive with file_data.
A base64-encoded data URL of the image (data:image/jpeg;base64,...). Mutually exclusive with file_id.
An optional filename. Only valid together with file_data.
Possible values: user
The role of the messages author, in this case user.
An optional name for the participant. Provides the model information to differentiate between participants of the same role.
The contents of the assistant message.
Possible values: assistant
The role of the messages author, in this case assistant.
An optional name for the participant. Provides the model information to differentiate between participants of the same role.
(Beta) Set this to true to force the model to start its answer by the content of the supplied prefix in this assistant message.
You must set base_url="https://api.deepseek.com/beta" to use this feature.
(Beta) Used for the thinking mode in the Chat Prefix Completion feature as the input for the CoT in the last assistant message. When using this feature, the prefix parameter must be set to true.
DeepSeek extension. The server_tool_calls of an earlier response, passed back unchanged. See the Web Tools Guide.
Possible values: tool
The role of the messages author, in this case tool.
required
The contents of the tool message. Either a string, or an array of content parts (for image input). See the Vision guide for details.
oneOf
string
Minimum items: 1
oneOf
Possible values: text
The type of the content part, in this case text.
The text content.
Possible values: image_url
The type of the content part, in this case image_url.
image_url
object
required
Either an http(s) URL of the image (max 8192 characters) or a base64-encoded data URL (data:image/jpeg;base64,...). Supported formats: JPEG, PNG, GIF, and WebP.
Possible values: low, high, original, auto
Controls how the image is processed. low downsamples the image to 512x512 (faster, cheaper). high, original, and auto keep the original image.
Exactly one of file_id and file_data is required: passing neither returns a 400 error ("file must have a file_id or file_data"); passing both returns a 400 error ("file cannot have both file_id and file_data").
Possible values: file
The type of the content part, in this case file.
The ID of a file uploaded via the Files API, of the form file-api-.... Mutually exclusive with file_data.
A base64-encoded data URL of the image (data:image/jpeg;base64,...). Mutually exclusive with file_id.
An optional filename. Only valid together with file_data.
Tool call that this message is responding to.
DeepSeek extension. Gives the model the current date, location, platform and language, like DeepSeek Chat. Place it directly before the user message it belongs to; anywhere else it returns 400. See the User Metadata Marker Guide.
Possible values: latest_reminder
The role of the messages author, in this case latest_reminder.
<date>,<weekday>,<location>,<platform>,<language>, for example 2026-11-11,Wednesday,Singapore,Web,English.
Possible values: deepseek-flash, deepseek-v4-pro
ID of the model to use. Use deepseek-flash or deepseek-v4-pro.
thinking
object
nullable
Controls the switch between thinking and non-thinking mode. When thinking is omitted or null, the mode follows reasoning_effort; if reasoning_effort is also omitted, thinking is enabled. When this object is provided, type is required and takes precedence over reasoning_effort for selecting the mode.
Possible values: enabled, disabled
If set to enabled, then use thinking mode. If set to disabled, then use non-thinking model.
Possible values: none, low, high, max
Controls the thinking mode toggle and the thinking effort. none disables thinking mode; low / high / max enable thinking mode. The default effort is high. For compatibility with existing software, minimal is accepted and mapped to low, and medium / xhigh are accepted and mapped to high.
The maximum number of tokens that can be generated in the chat completion.
The total length of input tokens and generated tokens is limited by the model's context length.
The value must be between 1 and 384K (393216). When not set, the default is 8K in non-thinking mode, 64K in thinking mode (128K with reasoning_effort set to max). Please refer to the Models & Pricing page for details.
response_format
object
nullable
An object specifying the format that the model must output. If omitted or null, plain text is used. When this object is provided, type is required.
Setting to { "type": "json_object" } enables JSON Output, which guarantees the message the model generates is valid JSON.
Important: When using JSON Output, you must also instruct the model to produce JSON yourself via a system or user message. If the model's output cannot be recognized as JSON, the API replaces the non-JSON content with spaces (U+0020). As a result, the message content may contain only spaces (appearing empty), even when finish_reason is stop. Also note that the message content may be partially cut off if finish_reason="length", which indicates the generation exceeded max_tokens or the conversation exceeded the max context length.
Possible values: text, json_object
Must be one of text or json_object.
nullable
Default value: null
Up to 16 sequences where the API will stop generating further tokens. Each sequence must be a non-empty string of at most 512 bytes (UTF-8 encoded).
oneOf
string
Minimum items: 1
Maximum items: 16
string
Default value: false
If set, partial message deltas will be sent. Tokens will be sent as data-only server-sent events (SSE) as they become available, with the stream terminated by a data: [DONE] message.
stream_options
object
nullable
Default value: null
Options for streaming response. Must be set together with stream: true; if stream is not set to true, the API returns a 400 error.
If set to true, all chunks in the stream will include a usage field, whose value is null on every chunk except the last one. If omitted or set to false, the usage field is absent from all chunks except the last one.
Either way, the last chunk before the data: [DONE] message carries the token usage statistics for the entire request in its usage field. Note that no separate usage-only chunk is emitted: the statistics ride on the last content chunk, whose choices array always contains exactly one element that carries no new content and a non-null finish_reason.
Possible values: >= 0 and <= 2
Default value: 1
What sampling temperature to use, between 0 and 2. Higher values like 0.8 will make the output more random, while lower values like 0.2 will make it more focused and deterministic.
We generally recommend altering this or top_p but not both. Has no effect in thinking mode.
Possible values: > 0 and <= 1
An alternative to sampling with temperature, called nucleus sampling, where the model considers the results of the tokens with top_p probability mass. So 0.1 means only the tokens comprising the top 10% probability mass are considered.
The value must be greater than 0 and at most 1. We generally recommend altering this or temperature but not both. The effective value depends on the model and mode:
tools
object[]
nullable
A list of tools the model may call: functions, and the web tools deepseek_web_search and deepseek_read_url (see the Web Tools Guide).
Use functions to provide a list of functions the model may generate JSON inputs for. Tool names must be unique.
oneOf
Possible values: function
The type of the tool, in this case function.
function
object
required
A description of what the function does, used by the model to choose when and how to call the function.
The name of the function to be called. Must be a-z, A-Z, 0-9, or contain underscores and dashes, with a maximum length of 128.
parameters
object
The parameters the functions accepts, described as a JSON Schema object. See the Tool Calls Guide for examples, and the JSON Schema reference for documentation about the format.
Omitting parameters defines a function with an empty parameter list.
Default value: false
If set to true, the API will use strict-mode for the tool calls to ensure the output always complies with the function's JSON schema. This is a Beta feature, for more details please refer to Tool Calls Guide
DeepSeek extension. Lets the model search the web; with thinking enabled, it runs deep thinking search. See the Web Tools Guide.
Possible values: deepseek_web_search
The type of the tool, in this case deepseek_web_search.
DeepSeek extension. Lets the model read the links in the last user message. See the Web Tools Guide.
Possible values: deepseek_read_url
The type of the tool, in this case deepseek_read_url.
nullable
Controls which (if any) tool is called by the model.
none means the model will not call any tool and instead generates a message.
auto means the model can pick between generating a message or calling one or more tools.
required means the model must call one or more tools.
Specifying a particular tool via {"type": "function", "function": {"name": "my_function"}} forces the model to call that tool.
none is the default when no tools are present. auto is the default if tools are present.
required and named tool choices are not supported in thinking mode; the API returns a 400 error. Disable thinking mode first to use them. With the web tools declared, required is treated as auto.
oneOf
string
Possible values: none, auto, required
Specifies a tool the model should use. Use to force the model to call a specific function.
Possible values: function
The type of the tool. Currently, only function is supported.
function
object
required
The name of the function to call.
safety_check
object
nullable
DeepSeek extension. Set level to high to turn on basic risk control, which reviews the model input and output. See the Model Content Safety Guide.
Possible values: low, high
high: basic risk control. low: the default checks, the same as omitting safety_check.
title
object
nullable
DeepSeek extension. Set {"enabled": true} to have the response carry a short title for the conversation. See the Conversation Title Guide.
Whether to return a conversation title.
Default value: false
Whether to return log probabilities of the output tokens or not. If true, returns the log probabilities of each output token returned in the content of message.
Possible values: >= 0 and <= 20
Default value: null
An integer between 0 and 20 specifying the number of most likely tokens to return at each token position, each with an associated log probability. logprobs must be set to true if this parameter is used.
A custom user_id. We recommend using only ASCII letters, digits, underscores (_), and hyphens (-). The UTF-8 encoded value must not exceed 512 bytes. Do not include user privacy information in the user_id.
This parameter is no longer supported. It will not take effect if you pass it to the API.
This parameter is no longer supported. It will not take effect if you pass it to the API.
Responses
- 200 (No streaming)
- 200 (Streaming)
OK, returns a chat completion object
- application/json
- Schema
- Example
Schema
Array [
Array [
]
Array [
Array [
]
]
Array [
Array [
]
]
]
A unique identifier for the chat completion.
choices
object[]
required
A list of chat completion choices.
Possible values: stop, length, content_filter, tool_calls, insufficient_system_resource, aborted
The reason the model stopped generating tokens. This will be stop if the model hit a natural stop point or a provided stop sequence,
length if the maximum number of tokens specified in the request was reached,
content_filter if content was omitted due to a flag from our content filters,
tool_calls if the model called a tool,
insufficient_system_resource if the request is interrupted due to insufficient resource of the inference system,
or aborted if the generation was interrupted.
The index of the choice in the list of choices.
message
object
required
A chat completion message generated by the model.
The contents of the message.
For thinking mode only. The reasoning contents of the assistant message, before the final answer.
tool_calls
object[]
The tool calls generated by the model.
The ID of the tool call.
Possible values: function
The type of the tool. Currently, only function is supported.
function
object
required
The function that the model called.
The name of the function to call.
The arguments to call the function with, as generated by the model in JSON format. Note that the model does not always generate valid JSON, and may hallucinate parameters not defined by your function schema. Validate the arguments in your code before calling your function.
Possible values: assistant
The role of the author of this message.
With the web tools. The citations in content. See Web Tools - Citations.
DeepSeek extension. With the web tools. The server-side tool calls of this response, one entry per step; pass it back unchanged in the next request. See the Web Tools Guide.
DeepSeek extension. The conversation title, returned when title.enabled is true. See the Conversation Title Guide.
Names the blocked images when images in the request fail the model content safety review. See the Model Content Safety Guide.
logprobs
object
nullable
required
Log probability information for the choice.
content
object[]
nullable
required
A list of message content tokens with log probability information.
The token.
The log probability of this token, if it is within the top 20 most likely tokens. Otherwise, the value -9999.0 is used to signify that the token is very unlikely.
A list of integers representing the UTF-8 bytes representation of the token. Useful in instances where characters are represented by multiple tokens and their byte representations must be combined to generate the correct text representation. Can be null if there is no bytes representation for the token.
top_logprobs
object[]
required
List of the most likely tokens and their log probability, at this token position. In rare cases, there may be fewer than the number of requested top_logprobs returned.
The token.
The log probability of this token, if it is within the top 20 most likely tokens. Otherwise, the value -9999.0 is used to signify that the token is very unlikely.
A list of integers representing the UTF-8 bytes representation of the token. Useful in instances where characters are represented by multiple tokens and their byte representations must be combined to generate the correct text representation. Can be null if there is no bytes representation for the token.
reasoning_content
object[]
nullable
A list of message content tokens with log probability information.
The token.
The log probability of this token, if it is within the top 20 most likely tokens. Otherwise, the value -9999.0 is used to signify that the token is very unlikely.
A list of integers representing the UTF-8 bytes representation of the token. Useful in instances where characters are represented by multiple tokens and their byte representations must be combined to generate the correct text representation. Can be null if there is no bytes representation for the token.
top_logprobs
object[]
required
List of the most likely tokens and their log probability, at this token position. In rare cases, there may be fewer than the number of requested top_logprobs returned.
The token.
The log probability of this token, if it is within the top 20 most likely tokens. Otherwise, the value -9999.0 is used to signify that the token is very unlikely.
A list of integers representing the UTF-8 bytes representation of the token. Useful in instances where characters are represented by multiple tokens and their byte representations must be combined to generate the correct text representation. Can be null if there is no bytes representation for the token.
The Unix timestamp (in seconds) of when the chat completion was created.
The model used for the chat completion.
This fingerprint represents the backend configuration that the model runs with.
Possible values: chat.completion
The object type, which is always chat.completion.
usage
object
Usage statistics for the completion request.
Number of tokens in the generated completion.
Number of tokens in the prompt. It equals prompt_cache_hit_tokens + prompt_cache_miss_tokens.
prompt_tokens_details
object
required
Breakdown of tokens used in the prompt.
Number of tokens in the prompt that hit the context cache. Same as prompt_cache_hit_tokens.
Number of tokens in the prompt that hits the context cache.
Number of tokens in the prompt that misses the context cache.
Total number of tokens used in the request (prompt + completion).
completion_tokens_details
object
Breakdown of tokens used in a completion.
Tokens generated by the model for reasoning.
Possible values: >= 0
DeepSeek extension. Tokens of the safety review, already included in completion_tokens. Present only when safety_check is high.
server_tool_use
object
With the web tools. The numbers of searches and fetches; omitted when all counts are zero. See the Web Tools Guide.
Possible values: >= 0
Number of searches, excluding image searches.
Possible values: >= 0
Number of page fetches.
Possible values: >= 0
Number of image searches.
{
"id": "88beaa90-17fd-45fc-939c-2ef75e2d913c",
"object": "chat.completion",
"created": 1793084842,
"model": "deepseek-flash",
"choices": [
{
"index": 0,
"message": {
"role": "assistant",
"content": "Hello! 👋 How can I help you today?"
},
"logprobs": null,
"finish_reason": "stop"
}
],
"usage": {
"prompt_tokens": 7,
"completion_tokens": 11,
"total_tokens": 18,
"prompt_tokens_details": {
"cached_tokens": 0
},
"prompt_cache_hit_tokens": 0,
"prompt_cache_miss_tokens": 7
},
"system_fingerprint": "e72d3c7f663d3e1383b00b3884cb4501"
}
OK, returns a streamed sequence of chat completion chunk objects
- text/event-stream
- Schema
- Example
Schema
Array [
Array [
Array [
]
Array [
Array [
]
]
Array [
Array [
]
]
]
]
Represents a streamed chunk of a chat completion response returned by model, based on the provided input.
A unique identifier for the chat completion. Each chunk has the same ID.
choices
object[]
required
A list of chat completion choices.
delta
object
required
A chat completion delta generated by streamed model responses.
The contents of the chunk message.
For thinking mode only. The reasoning contents of the assistant message, before the final answer.
Possible values: assistant
The role of the author of this message.
tool_calls
object[]
The tool calls generated by the model, such as function calls. The first chunk of each tool call carries the id, type and function fields; subsequent chunks only carry the function arguments.
The ID of the tool call.
Possible values: function
The type of the tool. Currently, only function is supported.
function
object
The name of the function to call.
The arguments to call the function with, as generated by the model in JSON format. Note that the model does not always generate valid JSON, and may hallucinate parameters not defined by your function schema. Validate the arguments in your code before calling your function.
With the web tools. The citations in content. See Web Tools - Citations. In streaming, sent once in the last chunk.
DeepSeek extension. With the web tools. The server-side tool calls of this response, one entry per step; pass it back unchanged in the next request. See the Web Tools Guide. In streaming, sent once in the last chunk.
DeepSeek extension. The conversation title, returned when title.enabled is true. See the Conversation Title Guide. In streaming, sent once in the last chunk.
Names the blocked images when images in the request fail the model content safety review. See the Model Content Safety Guide. In streaming, sent once in the last chunk.
DeepSeek extension. With the web tools. The progress of a server-side tool call; a chunk carries at most one event. See the Web Tools Guide.
logprobs
object
nullable
Log probability information for the choice.
content
object[]
nullable
required
A list of message content tokens with log probability information.
The token.
The log probability of this token, if it is within the top 20 most likely tokens. Otherwise, the value -9999.0 is used to signify that the token is very unlikely.
A list of integers representing the UTF-8 bytes representation of the token. Useful in instances where characters are represented by multiple tokens and their byte representations must be combined to generate the correct text representation. Can be null if there is no bytes representation for the token.
top_logprobs
object[]
required
List of the most likely tokens and their log probability, at this token position. In rare cases, there may be fewer than the number of requested top_logprobs returned.
The token.
The log probability of this token, if it is within the top 20 most likely tokens. Otherwise, the value -9999.0 is used to signify that the token is very unlikely.
A list of integers representing the UTF-8 bytes representation of the token. Useful in instances where characters are represented by multiple tokens and their byte representations must be combined to generate the correct text representation. Can be null if there is no bytes representation for the token.
reasoning_content
object[]
nullable
A list of message content tokens with log probability information.
The token.
The log probability of this token, if it is within the top 20 most likely tokens. Otherwise, the value -9999.0 is used to signify that the token is very unlikely.
A list of integers representing the UTF-8 bytes representation of the token. Useful in instances where characters are represented by multiple tokens and their byte representations must be combined to generate the correct text representation. Can be null if there is no bytes representation for the token.
top_logprobs
object[]
required
List of the most likely tokens and their log probability, at this token position. In rare cases, there may be fewer than the number of requested top_logprobs returned.
The token.
The log probability of this token, if it is within the top 20 most likely tokens. Otherwise, the value -9999.0 is used to signify that the token is very unlikely.
A list of integers representing the UTF-8 bytes representation of the token. Useful in instances where characters are represented by multiple tokens and their byte representations must be combined to generate the correct text representation. Can be null if there is no bytes representation for the token.
Possible values: stop, length, content_filter, tool_calls, insufficient_system_resource, aborted
The reason the model stopped generating tokens. This will be stop if the model hit a natural stop point or a provided stop sequence,
length if the maximum number of tokens specified in the request was reached,
content_filter if content was omitted due to a flag from our content filters,
tool_calls if the model called a tool,
insufficient_system_resource if the request is interrupted due to insufficient resource of the inference system,
or aborted if the generation was interrupted.
The index of the choice in the list of choices.
The Unix timestamp (in seconds) of when the chat completion was created. Each chunk has the same timestamp.
The model to generate the completion.
This fingerprint represents the backend configuration that the model runs with.
Possible values: chat.completion.chunk
The object type, which is always chat.completion.chunk.
data: {"id":"071f229d-daa0-4557-a580-106a22e8df73","object":"chat.completion.chunk","created":1793084907,"model":"deepseek-flash","system_fingerprint":"e72d3c7f663d3e1383b00b3884cb4501","choices":[{"index":0,"delta":{"role":"assistant","content":""},"logprobs":null,"finish_reason":null}],"usage":null}
data: {"id":"071f229d-daa0-4557-a580-106a22e8df73","object":"chat.completion.chunk","created":1793084907,"model":"deepseek-flash","system_fingerprint":"e72d3c7f663d3e1383b00b3884cb4501","choices":[{"index":0,"delta":{"content":"Hello"},"logprobs":null,"finish_reason":null}],"usage":null}
data: {"id":"071f229d-daa0-4557-a580-106a22e8df73","object":"chat.completion.chunk","created":1793084907,"model":"deepseek-flash","system_fingerprint":"e72d3c7f663d3e1383b00b3884cb4501","choices":[{"index":0,"delta":{"content":"!"},"logprobs":null,"finish_reason":null}],"usage":null}
data: {"id":"071f229d-daa0-4557-a580-106a22e8df73","object":"chat.completion.chunk","created":1793084907,"model":"deepseek-flash","system_fingerprint":"e72d3c7f663d3e1383b00b3884cb4501","choices":[{"index":0,"delta":{"content":" 👋"},"logprobs":null,"finish_reason":null}],"usage":null}
data: {"id":"071f229d-daa0-4557-a580-106a22e8df73","object":"chat.completion.chunk","created":1793084907,"model":"deepseek-flash","system_fingerprint":"e72d3c7f663d3e1383b00b3884cb4501","choices":[{"index":0,"delta":{"content":" How"},"logprobs":null,"finish_reason":null}],"usage":null}
data: {"id":"071f229d-daa0-4557-a580-106a22e8df73","object":"chat.completion.chunk","created":1793084907,"model":"deepseek-flash","system_fingerprint":"e72d3c7f663d3e1383b00b3884cb4501","choices":[{"index":0,"delta":{"content":" can"},"logprobs":null,"finish_reason":null}],"usage":null}
data: {"id":"071f229d-daa0-4557-a580-106a22e8df73","object":"chat.completion.chunk","created":1793084907,"model":"deepseek-flash","system_fingerprint":"e72d3c7f663d3e1383b00b3884cb4501","choices":[{"index":0,"delta":{"content":" I"},"logprobs":null,"finish_reason":null}],"usage":null}
data: {"id":"071f229d-daa0-4557-a580-106a22e8df73","object":"chat.completion.chunk","created":1793084907,"model":"deepseek-flash","system_fingerprint":"e72d3c7f663d3e1383b00b3884cb4501","choices":[{"index":0,"delta":{"content":" help"},"logprobs":null,"finish_reason":null}],"usage":null}
data: {"id":"071f229d-daa0-4557-a580-106a22e8df73","object":"chat.completion.chunk","created":1793084907,"model":"deepseek-flash","system_fingerprint":"e72d3c7f663d3e1383b00b3884cb4501","choices":[{"index":0,"delta":{"content":" you"},"logprobs":null,"finish_reason":null}],"usage":null}
data: {"id":"071f229d-daa0-4557-a580-106a22e8df73","object":"chat.completion.chunk","created":1793084907,"model":"deepseek-flash","system_fingerprint":"e72d3c7f663d3e1383b00b3884cb4501","choices":[{"index":0,"delta":{"content":" today"},"logprobs":null,"finish_reason":null}],"usage":null}
data: {"id":"071f229d-daa0-4557-a580-106a22e8df73","object":"chat.completion.chunk","created":1793084907,"model":"deepseek-flash","system_fingerprint":"e72d3c7f663d3e1383b00b3884cb4501","choices":[{"index":0,"delta":{"content":"?"},"logprobs":null,"finish_reason":null}],"usage":null}
data: {"id":"071f229d-daa0-4557-a580-106a22e8df73","object":"chat.completion.chunk","created":1793084907,"model":"deepseek-flash","system_fingerprint":"e72d3c7f663d3e1383b00b3884cb4501","choices":[{"index":0,"delta":{"content":""},"logprobs":null,"finish_reason":"stop"}],"usage":{"prompt_tokens":7,"completion_tokens":11,"total_tokens":18,"prompt_tokens_details":{"cached_tokens":0},"prompt_cache_hit_tokens":0,"prompt_cache_miss_tokens":7}}
data: [DONE]