Batches
Cancel a Message Batch
Create a Message Batch
Delete a Message Batch
List Message Batches
Retrieve Message Batch results
Retrieve a Message Batch
ModelsExpand Collapse
DeletedMessageBatch = object { id, type }
ID of the Message Batch.
type: "message_batch_deleted"Deleted object type.
For Message Batches, this is always "message_batch_deleted".
Deleted object type.
For Message Batches, this is always "message_batch_deleted".
MessageBatch = object { id, archived_at, cancel_initiated_at, 7 more }
Unique object identifier.
The format and length of IDs may change over time.
RFC 3339 datetime string representing the time at which the Message Batch was archived and its results became unavailable.
RFC 3339 datetime string representing the time at which cancellation was initiated for the Message Batch. Specified only if cancellation was initiated.
RFC 3339 datetime string representing the time at which the Message Batch was created.
RFC 3339 datetime string representing the time at which processing for the Message Batch ended. Specified only once processing ends.
Processing ends when every request in a Message Batch has either succeeded, errored, canceled, or expired.
RFC 3339 datetime string representing the time at which the Message Batch will expire and end processing, which is 24 hours after creation.
processing_status: "in_progress" or "canceling" or "ended"Processing status of the Message Batch.
Processing status of the Message Batch.
Tallies requests within the Message Batch, categorized by their status.
Requests start as processing and move to one of the other statuses only once processing of the entire batch ends. The sum of all values always matches the total number of requests in the batch.
Tallies requests within the Message Batch, categorized by their status.
Requests start as processing and move to one of the other statuses only once processing of the entire batch ends. The sum of all values always matches the total number of requests in the batch.
Number of requests in the Message Batch that have been canceled.
This is zero until processing of the entire Message Batch has ended.
Number of requests in the Message Batch that encountered an error.
This is zero until processing of the entire Message Batch has ended.
Number of requests in the Message Batch that have expired.
This is zero until processing of the entire Message Batch has ended.
Number of requests in the Message Batch that are processing.
Number of requests in the Message Batch that have completed successfully.
This is zero until processing of the entire Message Batch has ended.
URL to a .jsonl file containing the results of the Message Batch requests. Specified only once processing ends.
Results in the file are not guaranteed to be in the same order as requests. Use the custom_id field to match results to requests.
type: "message_batch"Object type.
For Message Batches, this is always "message_batch".
Object type.
For Message Batches, this is always "message_batch".
MessageBatchCanceledResult = object { type }
type: "canceled"
MessageBatchErroredResult = object { error, type }
InvalidRequestError = object { message, type }
type: "invalid_request_error"
AuthenticationError = object { message, type }
type: "authentication_error"
BillingError = object { message, type }
type: "billing_error"
PermissionError = object { message, type }
type: "permission_error"
NotFoundError = object { message, type }
type: "not_found_error"
RateLimitError = object { message, type }
type: "rate_limit_error"
GatewayTimeoutError = object { message, type }
type: "timeout_error"
APIErrorObject = object { message, type }
type: "api_error"
OverloadedError = object { message, type }
type: "overloaded_error"
type: "error"
type: "errored"
MessageBatchExpiredResult = object { type }
type: "expired"
MessageBatchIndividualResponse = object { custom_id, result } This is a single line in the response .jsonl file and does not represent the response as a whole.
This is a single line in the response .jsonl file and does not represent the response as a whole.
Developer-provided ID created for each request in a Message Batch. Useful for matching results to requests, as results may be given out of request order.
Must be unique for each request within the Message Batch.
Processing result for this request.
Contains a Message output if processing was successful, an error response if processing failed, or the reason why processing was not attempted, such as cancellation or expiration.
Processing result for this request.
Contains a Message output if processing was successful, an error response if processing failed, or the reason why processing was not attempted, such as cancellation or expiration.
MessageBatchSucceededResult = object { message, type }
Unique object identifier.
The format and length of IDs may change over time.
Content generated by the model.
This is an array of content blocks, each of which has a type that determines its shape.
Example:
[{"type": "text", "text": "Hi, I'm Claude."}]
If the request input messages ended with an assistant turn, then the response content will continue directly from that last turn. You can use this to constrain the model's output.
For example, if the input messages were:
[
{"role": "user", "content": "What's the Greek name for Sun? (A) Sol (B) Helios (C) Sun"},
{"role": "assistant", "content": "The best answer is ("}
]
Then the response content might be:
[{"type": "text", "text": "B)"}]
Content generated by the model.
This is an array of content blocks, each of which has a type that determines its shape.
Example:
[{"type": "text", "text": "Hi, I'm Claude."}]
If the request input messages ended with an assistant turn, then the response content will continue directly from that last turn. You can use this to constrain the model's output.
For example, if the input messages were:
[
{"role": "user", "content": "What's the Greek name for Sun? (A) Sol (B) Helios (C) Sun"},
{"role": "assistant", "content": "The best answer is ("}
]
Then the response content might be:
[{"type": "text", "text": "B)"}]
TextBlock = object { citations, text, type }
Citations supporting the text block.
The type of citation returned will depend on the type of document being cited. Citing a PDF results in page_location, plain text results in char_location, and content document results in content_block_location.
Citations supporting the text block.
The type of citation returned will depend on the type of document being cited. Citing a PDF results in page_location, plain text results in char_location, and content document results in content_block_location.
CitationCharLocation = object { cited_text, document_index, document_title, 4 more }
type: "char_location"
CitationPageLocation = object { cited_text, document_index, document_title, 4 more }
type: "page_location"
CitationContentBlockLocation = object { cited_text, document_index, document_title, 4 more }
type: "content_block_location"
CitationsWebSearchResultLocation = object { cited_text, encrypted_index, title, 2 more }
type: "web_search_result_location"
CitationsSearchResultLocation = object { cited_text, end_block_index, search_result_index, 4 more }
type: "search_result_location"
type: "text"
ThinkingBlock = object { signature, thinking, type }
type: "thinking"
RedactedThinkingBlock = object { data, type }
type: "redacted_thinking"
ToolUseBlock = object { id, input, name, type }
type: "tool_use"
ServerToolUseBlock = object { id, input, name, type }
name: "web_search"
type: "server_tool_use"
WebSearchToolResultBlock = object { content, tool_use_id, type }
WebSearchToolResultError = object { error_code, type }
error_code: "invalid_tool_input" or "unavailable" or "max_uses_exceeded" or 2 more
type: "web_search_tool_result_error"
type: "web_search_result"
type: "web_search_tool_result"
UnionMember0 = "claude-3-7-sonnet-latest" or "claude-3-7-sonnet-20250219" or "claude-3-5-haiku-latest" or 15 moreThe model that will complete your prompt.
See models for additional details and options.
The model that will complete your prompt.
See models for additional details and options.
High-performance model with early extended thinking
High-performance model with early extended thinking
Fastest and most compact model for near-instant responsiveness
Our fastest model
Hybrid model, capable of near-instant responses and extended thinking
Hybrid model, capable of near-instant responses and extended thinking
High-performance model with extended thinking
High-performance model with extended thinking
High-performance model with extended thinking
Our best model for real-world agents and coding
Our best model for real-world agents and coding
Our most capable model
Our most capable model
Our most capable model
Our most capable model
Excels at writing and complex tasks
Excels at writing and complex tasks
Our previous most fast and cost-effective
role: "assistant"Conversational role of the generated message.
This will always be "assistant".
Conversational role of the generated message.
This will always be "assistant".
The reason that we stopped.
This may be one the following values:
"end_turn": the model reached a natural stopping point"max_tokens": we exceeded the requested max_tokens or the model's maximum"stop_sequence": one of your provided custom stop_sequences was generated"tool_use": the model invoked one or more tools"pause_turn": we paused a long-running turn. You may provide the response back as-is in a subsequent request to let the model continue."refusal": when streaming classifiers intervene to handle potential policy violations
In non-streaming mode this value is always non-null. In streaming mode, it is null in the message_start event and non-null otherwise.
The reason that we stopped.
This may be one the following values:
"end_turn": the model reached a natural stopping point"max_tokens": we exceeded the requestedmax_tokensor the model's maximum"stop_sequence": one of your provided customstop_sequenceswas generated"tool_use": the model invoked one or more tools"pause_turn": we paused a long-running turn. You may provide the response back as-is in a subsequent request to let the model continue."refusal": when streaming classifiers intervene to handle potential policy violations
In non-streaming mode this value is always non-null. In streaming mode, it is null in the message_start event and non-null otherwise.
Which custom stop sequence was generated, if any.
This value will be a non-null string if one of your custom stop sequences was generated.
type: "message"Object type.
For Messages, this is always "message".
Object type.
For Messages, this is always "message".
Billing and rate-limit usage.
Anthropic's API bills and rate-limits by token counts, as tokens represent the underlying cost to our systems.
Under the hood, the API transforms requests into a format suitable for the model. The model's output then goes through a parsing stage before becoming an API response. As a result, the token counts in usage will not match one-to-one with the exact visible content of an API request or response.
For example, output_tokens will be non-zero, even for an empty string response from Claude.
Total input tokens in a request is the summation of input_tokens, cache_creation_input_tokens, and cache_read_input_tokens.
Billing and rate-limit usage.
Anthropic's API bills and rate-limits by token counts, as tokens represent the underlying cost to our systems.
Under the hood, the API transforms requests into a format suitable for the model. The model's output then goes through a parsing stage before becoming an API response. As a result, the token counts in usage will not match one-to-one with the exact visible content of an API request or response.
For example, output_tokens will be non-zero, even for an empty string response from Claude.
Total input tokens in a request is the summation of input_tokens, cache_creation_input_tokens, and cache_read_input_tokens.
Breakdown of cached tokens by TTL
Breakdown of cached tokens by TTL
The number of input tokens used to create the 1 hour cache entry.
The number of input tokens used to create the 5 minute cache entry.
The number of input tokens used to create the cache entry.
The number of input tokens read from the cache.
The number of input tokens which were used.
The number of output tokens which were used.
The number of server tool requests.
The number of server tool requests.
The number of web search tool requests.
service_tier: "standard" or "priority" or "batch"If the request used the priority, standard, or batch tier.
If the request used the priority, standard, or batch tier.
type: "succeeded"
MessageBatchErroredResult = object { error, type }
InvalidRequestError = object { message, type }
type: "invalid_request_error"
AuthenticationError = object { message, type }
type: "authentication_error"
BillingError = object { message, type }
type: "billing_error"
PermissionError = object { message, type }
type: "permission_error"
NotFoundError = object { message, type }
type: "not_found_error"
RateLimitError = object { message, type }
type: "rate_limit_error"
GatewayTimeoutError = object { message, type }
type: "timeout_error"
APIErrorObject = object { message, type }
type: "api_error"
OverloadedError = object { message, type }
type: "overloaded_error"
type: "error"
type: "errored"
MessageBatchCanceledResult = object { type }
type: "canceled"
MessageBatchExpiredResult = object { type }
type: "expired"
MessageBatchRequestCounts = object { canceled, errored, expired, 2 more }
Number of requests in the Message Batch that have been canceled.
This is zero until processing of the entire Message Batch has ended.
Number of requests in the Message Batch that encountered an error.
This is zero until processing of the entire Message Batch has ended.
Number of requests in the Message Batch that have expired.
This is zero until processing of the entire Message Batch has ended.
Number of requests in the Message Batch that are processing.
Number of requests in the Message Batch that have completed successfully.
This is zero until processing of the entire Message Batch has ended.
MessageBatchResult = MessageBatchSucceededResult { message, type } or MessageBatchErroredResult { error, type } or MessageBatchCanceledResult { type } or MessageBatchExpiredResult { type } Processing result for this request.
Contains a Message output if processing was successful, an error response if processing failed, or the reason why processing was not attempted, such as cancellation or expiration.
Processing result for this request.
Contains a Message output if processing was successful, an error response if processing failed, or the reason why processing was not attempted, such as cancellation or expiration.
MessageBatchSucceededResult = object { message, type }
Unique object identifier.
The format and length of IDs may change over time.
Content generated by the model.
This is an array of content blocks, each of which has a type that determines its shape.
Example:
[{"type": "text", "text": "Hi, I'm Claude."}]
If the request input messages ended with an assistant turn, then the response content will continue directly from that last turn. You can use this to constrain the model's output.
For example, if the input messages were:
[
{"role": "user", "content": "What's the Greek name for Sun? (A) Sol (B) Helios (C) Sun"},
{"role": "assistant", "content": "The best answer is ("}
]
Then the response content might be:
[{"type": "text", "text": "B)"}]
Content generated by the model.
This is an array of content blocks, each of which has a type that determines its shape.
Example:
[{"type": "text", "text": "Hi, I'm Claude."}]
If the request input messages ended with an assistant turn, then the response content will continue directly from that last turn. You can use this to constrain the model's output.
For example, if the input messages were:
[
{"role": "user", "content": "What's the Greek name for Sun? (A) Sol (B) Helios (C) Sun"},
{"role": "assistant", "content": "The best answer is ("}
]
Then the response content might be:
[{"type": "text", "text": "B)"}]
TextBlock = object { citations, text, type }
Citations supporting the text block.
The type of citation returned will depend on the type of document being cited. Citing a PDF results in page_location, plain text results in char_location, and content document results in content_block_location.
Citations supporting the text block.
The type of citation returned will depend on the type of document being cited. Citing a PDF results in page_location, plain text results in char_location, and content document results in content_block_location.
CitationCharLocation = object { cited_text, document_index, document_title, 4 more }
type: "char_location"
CitationPageLocation = object { cited_text, document_index, document_title, 4 more }
type: "page_location"
CitationContentBlockLocation = object { cited_text, document_index, document_title, 4 more }
type: "content_block_location"
CitationsWebSearchResultLocation = object { cited_text, encrypted_index, title, 2 more }
type: "web_search_result_location"
CitationsSearchResultLocation = object { cited_text, end_block_index, search_result_index, 4 more }
type: "search_result_location"
type: "text"
ThinkingBlock = object { signature, thinking, type }
type: "thinking"
RedactedThinkingBlock = object { data, type }
type: "redacted_thinking"
ToolUseBlock = object { id, input, name, type }
type: "tool_use"
ServerToolUseBlock = object { id, input, name, type }
name: "web_search"
type: "server_tool_use"
WebSearchToolResultBlock = object { content, tool_use_id, type }
WebSearchToolResultError = object { error_code, type }
error_code: "invalid_tool_input" or "unavailable" or "max_uses_exceeded" or 2 more
type: "web_search_tool_result_error"
type: "web_search_result"
type: "web_search_tool_result"
UnionMember0 = "claude-3-7-sonnet-latest" or "claude-3-7-sonnet-20250219" or "claude-3-5-haiku-latest" or 15 moreThe model that will complete your prompt.
See models for additional details and options.
The model that will complete your prompt.
See models for additional details and options.
High-performance model with early extended thinking
High-performance model with early extended thinking
Fastest and most compact model for near-instant responsiveness
Our fastest model
Hybrid model, capable of near-instant responses and extended thinking
Hybrid model, capable of near-instant responses and extended thinking
High-performance model with extended thinking
High-performance model with extended thinking
High-performance model with extended thinking
Our best model for real-world agents and coding
Our best model for real-world agents and coding
Our most capable model
Our most capable model
Our most capable model
Our most capable model
Excels at writing and complex tasks
Excels at writing and complex tasks
Our previous most fast and cost-effective
role: "assistant"Conversational role of the generated message.
This will always be "assistant".
Conversational role of the generated message.
This will always be "assistant".
The reason that we stopped.
This may be one the following values:
"end_turn": the model reached a natural stopping point"max_tokens": we exceeded the requested max_tokens or the model's maximum"stop_sequence": one of your provided custom stop_sequences was generated"tool_use": the model invoked one or more tools"pause_turn": we paused a long-running turn. You may provide the response back as-is in a subsequent request to let the model continue."refusal": when streaming classifiers intervene to handle potential policy violations
In non-streaming mode this value is always non-null. In streaming mode, it is null in the message_start event and non-null otherwise.
The reason that we stopped.
This may be one the following values:
"end_turn": the model reached a natural stopping point"max_tokens": we exceeded the requestedmax_tokensor the model's maximum"stop_sequence": one of your provided customstop_sequenceswas generated"tool_use": the model invoked one or more tools"pause_turn": we paused a long-running turn. You may provide the response back as-is in a subsequent request to let the model continue."refusal": when streaming classifiers intervene to handle potential policy violations
In non-streaming mode this value is always non-null. In streaming mode, it is null in the message_start event and non-null otherwise.
Which custom stop sequence was generated, if any.
This value will be a non-null string if one of your custom stop sequences was generated.
type: "message"Object type.
For Messages, this is always "message".
Object type.
For Messages, this is always "message".
Billing and rate-limit usage.
Anthropic's API bills and rate-limits by token counts, as tokens represent the underlying cost to our systems.
Under the hood, the API transforms requests into a format suitable for the model. The model's output then goes through a parsing stage before becoming an API response. As a result, the token counts in usage will not match one-to-one with the exact visible content of an API request or response.
For example, output_tokens will be non-zero, even for an empty string response from Claude.
Total input tokens in a request is the summation of input_tokens, cache_creation_input_tokens, and cache_read_input_tokens.
Billing and rate-limit usage.
Anthropic's API bills and rate-limits by token counts, as tokens represent the underlying cost to our systems.
Under the hood, the API transforms requests into a format suitable for the model. The model's output then goes through a parsing stage before becoming an API response. As a result, the token counts in usage will not match one-to-one with the exact visible content of an API request or response.
For example, output_tokens will be non-zero, even for an empty string response from Claude.
Total input tokens in a request is the summation of input_tokens, cache_creation_input_tokens, and cache_read_input_tokens.
Breakdown of cached tokens by TTL
Breakdown of cached tokens by TTL
The number of input tokens used to create the 1 hour cache entry.
The number of input tokens used to create the 5 minute cache entry.
The number of input tokens used to create the cache entry.
The number of input tokens read from the cache.
The number of input tokens which were used.
The number of output tokens which were used.
The number of server tool requests.
The number of server tool requests.
The number of web search tool requests.
service_tier: "standard" or "priority" or "batch"If the request used the priority, standard, or batch tier.
If the request used the priority, standard, or batch tier.
type: "succeeded"
MessageBatchErroredResult = object { error, type }
InvalidRequestError = object { message, type }
type: "invalid_request_error"
AuthenticationError = object { message, type }
type: "authentication_error"
BillingError = object { message, type }
type: "billing_error"
PermissionError = object { message, type }
type: "permission_error"
NotFoundError = object { message, type }
type: "not_found_error"
RateLimitError = object { message, type }
type: "rate_limit_error"
GatewayTimeoutError = object { message, type }
type: "timeout_error"
APIErrorObject = object { message, type }
type: "api_error"
OverloadedError = object { message, type }
type: "overloaded_error"
type: "error"
type: "errored"
MessageBatchCanceledResult = object { type }
type: "canceled"
MessageBatchExpiredResult = object { type }
type: "expired"
MessageBatchSucceededResult = object { message, type }
Unique object identifier.
The format and length of IDs may change over time.
Content generated by the model.
This is an array of content blocks, each of which has a type that determines its shape.
Example:
[{"type": "text", "text": "Hi, I'm Claude."}]
If the request input messages ended with an assistant turn, then the response content will continue directly from that last turn. You can use this to constrain the model's output.
For example, if the input messages were:
[
{"role": "user", "content": "What's the Greek name for Sun? (A) Sol (B) Helios (C) Sun"},
{"role": "assistant", "content": "The best answer is ("}
]
Then the response content might be:
[{"type": "text", "text": "B)"}]
Content generated by the model.
This is an array of content blocks, each of which has a type that determines its shape.
Example:
[{"type": "text", "text": "Hi, I'm Claude."}]
If the request input messages ended with an assistant turn, then the response content will continue directly from that last turn. You can use this to constrain the model's output.
For example, if the input messages were:
[
{"role": "user", "content": "What's the Greek name for Sun? (A) Sol (B) Helios (C) Sun"},
{"role": "assistant", "content": "The best answer is ("}
]
Then the response content might be:
[{"type": "text", "text": "B)"}]
TextBlock = object { citations, text, type }
Citations supporting the text block.
The type of citation returned will depend on the type of document being cited. Citing a PDF results in page_location, plain text results in char_location, and content document results in content_block_location.
Citations supporting the text block.
The type of citation returned will depend on the type of document being cited. Citing a PDF results in page_location, plain text results in char_location, and content document results in content_block_location.
CitationCharLocation = object { cited_text, document_index, document_title, 4 more }
type: "char_location"
CitationPageLocation = object { cited_text, document_index, document_title, 4 more }
type: "page_location"
CitationContentBlockLocation = object { cited_text, document_index, document_title, 4 more }
type: "content_block_location"
CitationsWebSearchResultLocation = object { cited_text, encrypted_index, title, 2 more }
type: "web_search_result_location"
CitationsSearchResultLocation = object { cited_text, end_block_index, search_result_index, 4 more }
type: "search_result_location"
type: "text"
ThinkingBlock = object { signature, thinking, type }
type: "thinking"
RedactedThinkingBlock = object { data, type }
type: "redacted_thinking"
ToolUseBlock = object { id, input, name, type }
type: "tool_use"
ServerToolUseBlock = object { id, input, name, type }
name: "web_search"
type: "server_tool_use"
WebSearchToolResultBlock = object { content, tool_use_id, type }
WebSearchToolResultError = object { error_code, type }
error_code: "invalid_tool_input" or "unavailable" or "max_uses_exceeded" or 2 more
type: "web_search_tool_result_error"
type: "web_search_result"
type: "web_search_tool_result"
UnionMember0 = "claude-3-7-sonnet-latest" or "claude-3-7-sonnet-20250219" or "claude-3-5-haiku-latest" or 15 moreThe model that will complete your prompt.
See models for additional details and options.
The model that will complete your prompt.
See models for additional details and options.
High-performance model with early extended thinking
High-performance model with early extended thinking
Fastest and most compact model for near-instant responsiveness
Our fastest model
Hybrid model, capable of near-instant responses and extended thinking
Hybrid model, capable of near-instant responses and extended thinking
High-performance model with extended thinking
High-performance model with extended thinking
High-performance model with extended thinking
Our best model for real-world agents and coding
Our best model for real-world agents and coding
Our most capable model
Our most capable model
Our most capable model
Our most capable model
Excels at writing and complex tasks
Excels at writing and complex tasks
Our previous most fast and cost-effective
role: "assistant"Conversational role of the generated message.
This will always be "assistant".
Conversational role of the generated message.
This will always be "assistant".
The reason that we stopped.
This may be one the following values:
"end_turn": the model reached a natural stopping point"max_tokens": we exceeded the requested max_tokens or the model's maximum"stop_sequence": one of your provided custom stop_sequences was generated"tool_use": the model invoked one or more tools"pause_turn": we paused a long-running turn. You may provide the response back as-is in a subsequent request to let the model continue."refusal": when streaming classifiers intervene to handle potential policy violations
In non-streaming mode this value is always non-null. In streaming mode, it is null in the message_start event and non-null otherwise.
The reason that we stopped.
This may be one the following values:
"end_turn": the model reached a natural stopping point"max_tokens": we exceeded the requestedmax_tokensor the model's maximum"stop_sequence": one of your provided customstop_sequenceswas generated"tool_use": the model invoked one or more tools"pause_turn": we paused a long-running turn. You may provide the response back as-is in a subsequent request to let the model continue."refusal": when streaming classifiers intervene to handle potential policy violations
In non-streaming mode this value is always non-null. In streaming mode, it is null in the message_start event and non-null otherwise.
Which custom stop sequence was generated, if any.
This value will be a non-null string if one of your custom stop sequences was generated.
type: "message"Object type.
For Messages, this is always "message".
Object type.
For Messages, this is always "message".
Billing and rate-limit usage.
Anthropic's API bills and rate-limits by token counts, as tokens represent the underlying cost to our systems.
Under the hood, the API transforms requests into a format suitable for the model. The model's output then goes through a parsing stage before becoming an API response. As a result, the token counts in usage will not match one-to-one with the exact visible content of an API request or response.
For example, output_tokens will be non-zero, even for an empty string response from Claude.
Total input tokens in a request is the summation of input_tokens, cache_creation_input_tokens, and cache_read_input_tokens.
Billing and rate-limit usage.
Anthropic's API bills and rate-limits by token counts, as tokens represent the underlying cost to our systems.
Under the hood, the API transforms requests into a format suitable for the model. The model's output then goes through a parsing stage before becoming an API response. As a result, the token counts in usage will not match one-to-one with the exact visible content of an API request or response.
For example, output_tokens will be non-zero, even for an empty string response from Claude.
Total input tokens in a request is the summation of input_tokens, cache_creation_input_tokens, and cache_read_input_tokens.
Breakdown of cached tokens by TTL
Breakdown of cached tokens by TTL
The number of input tokens used to create the 1 hour cache entry.
The number of input tokens used to create the 5 minute cache entry.
The number of input tokens used to create the cache entry.
The number of input tokens read from the cache.
The number of input tokens which were used.
The number of output tokens which were used.
The number of server tool requests.
The number of server tool requests.
The number of web search tool requests.
service_tier: "standard" or "priority" or "batch"If the request used the priority, standard, or batch tier.
If the request used the priority, standard, or batch tier.