Reference 2026-08-13 22:00 UTC to 2026-08-14 23:57 UTC

3862 for reasoning items returned by `POST /v1/responses` and WebSocket3862 for reasoning items returned by `POST /v1/responses` and WebSocket

3863 `response.create` requests.3863 `response.create` requests.

3864 3864 

3865 When streaming, use the completed reasoning item and its

3866 `encrypted_content` from the `response.output_item.done` event in

3867 subsequent requests. The `encrypted_content` in

3868 `response.output_item.added` may be incomplete. This is especially

3869 important when `store` is `false` or when using Zero Data Retention.

3870 

3865 - `status: optional "in_progress" or "completed" or "incomplete"`3871 - `status: optional "in_progress" or "completed" or "incomplete"`

3866 3872 

3867 The status of the item. One of `in_progress`, `completed`, or3873 The status of the item. One of `in_progress`, `completed`, or

4635 Unique identifier for the MCP tool call approval request.4641 Unique identifier for the MCP tool call approval request.

4636 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.4642 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.

4637 4643 

4638 - `error: optional string or null`4644 - `error: optional BetaMcpToolCallError or null`

4639 4645 

4640 The error from the tool call, if any.4646 The error from the tool call, if any.

4641 4647 

4648 - `McpProtocolError object { code, message, type }`

4649 

4650 - `code: number`

4651 

4652 - `message: string`

4653 

4654 - `type: "mcp_protocol_error"`

4655 

4656 - `"mcp_protocol_error"`

4657 

4658 - `McpToolExecutionError object { content, type }`

4659 

4660 - `content: unknown`

4661 

4662 - `type: "mcp_tool_execution_error"`

4663 

4664 - `"mcp_tool_execution_error"`

4665 

4666 - `HTTPError object { code, message, type }`

4667 

4668 - `code: number`

4669 

4670 - `message: string`

4671 

4672 - `type: "http_error"`

4673 

4674 - `"http_error"`

4675 

4642 - `output: optional string or null`4676 - `output: optional string or null`

4643 4677 

4644 The output from the tool call.4678 The output from the tool call.

5924 for reasoning items returned by `POST /v1/responses` and WebSocket5958 for reasoning items returned by `POST /v1/responses` and WebSocket

5925 `response.create` requests.5959 `response.create` requests.

5926 5960 

5961 When streaming, use the completed reasoning item and its

5962 `encrypted_content` from the `response.output_item.done` event in

5963 subsequent requests. The `encrypted_content` in

5964 `response.output_item.added` may be incomplete. This is especially

5965 important when `store` is `false` or when using Zero Data Retention.

5966 

5927 - `status: optional "in_progress" or "completed" or "incomplete"`5967 - `status: optional "in_progress" or "completed" or "incomplete"`

5928 5968 

5929 The status of the item. One of `in_progress`, `completed`, or5969 The status of the item. One of `in_progress`, `completed`, or

8801 Unique identifier for the MCP tool call approval request.8841 Unique identifier for the MCP tool call approval request.

8802 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.8842 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.

8803 8843 

8804 - `error: optional string or null`8844 - `error: optional BetaMcpToolCallError or null`

8805 8845 

8806 The error from the tool call, if any.8846 The error from the tool call, if any.

8807 8847 

3992 for reasoning items returned by `POST /v1/responses` and WebSocket3992 for reasoning items returned by `POST /v1/responses` and WebSocket

3993 `response.create` requests.3993 `response.create` requests.

3994 3994 

3995 When streaming, use the completed reasoning item and its

3996 `encrypted_content` from the `response.output_item.done` event in

3997 subsequent requests. The `encrypted_content` in

3998 `response.output_item.added` may be incomplete. This is especially

3999 important when `store` is `false` or when using Zero Data Retention.

4000 

3995 - `status: optional "in_progress" or "completed" or "incomplete"`4001 - `status: optional "in_progress" or "completed" or "incomplete"`

3996 4002 

3997 The status of the item. One of `in_progress`, `completed`, or4003 The status of the item. One of `in_progress`, `completed`, or

4765 Unique identifier for the MCP tool call approval request.4771 Unique identifier for the MCP tool call approval request.

4766 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.4772 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.

4767 4773 

4768 - `error: optional string or null`4774 - `error: optional BetaMcpToolCallError or null`

4769 4775 

4770 The error from the tool call, if any.4776 The error from the tool call, if any.

4771 4777 

4778 - `McpProtocolError object { code, message, type }`

4779 

4780 - `code: number`

4781 

4782 - `message: string`

4783 

4784 - `type: "mcp_protocol_error"`

4785 

4786 - `"mcp_protocol_error"`

4787 

4788 - `McpToolExecutionError object { content, type }`

4789 

4790 - `content: unknown`

4791 

4792 - `type: "mcp_tool_execution_error"`

4793 

4794 - `"mcp_tool_execution_error"`

4795 

4796 - `HTTPError object { code, message, type }`

4797 

4798 - `code: number`

4799 

4800 - `message: string`

4801 

4802 - `type: "http_error"`

4803 

4804 - `"http_error"`

4805 

4772 - `output: optional string or null`4806 - `output: optional string or null`

4773 4807 

4774 The output from the tool call.4808 The output from the tool call.

8917 for reasoning items returned by `POST /v1/responses` and WebSocket8951 for reasoning items returned by `POST /v1/responses` and WebSocket

8918 `response.create` requests.8952 `response.create` requests.

8919 8953 

8954 When streaming, use the completed reasoning item and its

8955 `encrypted_content` from the `response.output_item.done` event in

8956 subsequent requests. The `encrypted_content` in

8957 `response.output_item.added` may be incomplete. This is especially

8958 important when `store` is `false` or when using Zero Data Retention.

8959 

8920 - `status: optional "in_progress" or "completed" or "incomplete"`8960 - `status: optional "in_progress" or "completed" or "incomplete"`

8921 8961 

8922 The status of the item. One of `in_progress`, `completed`, or8962 The status of the item. One of `in_progress`, `completed`, or

9676 Unique identifier for the MCP tool call approval request.9716 Unique identifier for the MCP tool call approval request.

9677 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.9717 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.

9678 9718 

9679 - `error: optional string or null`9719 - `error: optional BetaMcpToolCallError or null`

9680 9720 

9681 The error from the tool call, if any.9721 The error from the tool call, if any.

9682 9722 

9723 - `McpProtocolError object { code, message, type }`

9724 

9725 - `code: number`

9726 

9727 - `message: string`

9728 

9729 - `type: "mcp_protocol_error"`

9730 

9731 - `"mcp_protocol_error"`

9732 

9733 - `McpToolExecutionError object { content, type }`

9734 

9735 - `content: unknown`

9736 

9737 - `type: "mcp_tool_execution_error"`

9738 

9739 - `"mcp_tool_execution_error"`

9740 

9741 - `HTTPError object { code, message, type }`

9742 

9743 - `code: number`

9744 

9745 - `message: string`

9746 

9747 - `type: "http_error"`

9748 

9749 - `"http_error"`

9750 

9683 - `output: optional string or null`9751 - `output: optional string or null`

9684 9752 

9685 The output from the tool call.9753 The output from the tool call.

3854 for reasoning items returned by `POST /v1/responses` and WebSocket3854 for reasoning items returned by `POST /v1/responses` and WebSocket

3855 `response.create` requests.3855 `response.create` requests.

3856 3856 

3857 When streaming, use the completed reasoning item and its

3858 `encrypted_content` from the `response.output_item.done` event in

3859 subsequent requests. The `encrypted_content` in

3860 `response.output_item.added` may be incomplete. This is especially

3861 important when `store` is `false` or when using Zero Data Retention.

3862 

3857 - `status: optional "in_progress" or "completed" or "incomplete"`3863 - `status: optional "in_progress" or "completed" or "incomplete"`

3858 3864 

3859 The status of the item. One of `in_progress`, `completed`, or3865 The status of the item. One of `in_progress`, `completed`, or

4627 Unique identifier for the MCP tool call approval request.4633 Unique identifier for the MCP tool call approval request.

4628 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.4634 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.

4629 4635 

4630 - `error: optional string or null`4636 - `error: optional BetaMcpToolCallError or null`

4631 4637 

4632 The error from the tool call, if any.4638 The error from the tool call, if any.

4633 4639 

4640 - `McpProtocolError object { code, message, type }`

4641 

4642 - `code: number`

4643 

4644 - `message: string`

4645 

4646 - `type: "mcp_protocol_error"`

4647 

4648 - `"mcp_protocol_error"`

4649 

4650 - `McpToolExecutionError object { content, type }`

4651 

4652 - `content: unknown`

4653 

4654 - `type: "mcp_tool_execution_error"`

4655 

4656 - `"mcp_tool_execution_error"`

4657 

4658 - `HTTPError object { code, message, type }`

4659 

4660 - `code: number`

4661 

4662 - `message: string`

4663 

4664 - `type: "http_error"`

4665 

4666 - `"http_error"`

4667 

4634 - `output: optional string or null`4668 - `output: optional string or null`

4635 4669 

4636 The output from the tool call.4670 The output from the tool call.

10579 for reasoning items returned by `POST /v1/responses` and WebSocket10613 for reasoning items returned by `POST /v1/responses` and WebSocket

10580 `response.create` requests.10614 `response.create` requests.

10581 10615 

10616 When streaming, use the completed reasoning item and its

10617 `encrypted_content` from the `response.output_item.done` event in

10618 subsequent requests. The `encrypted_content` in

10619 `response.output_item.added` may be incomplete. This is especially

10620 important when `store` is `false` or when using Zero Data Retention.

10621 

10582 - `status: optional "in_progress" or "completed" or "incomplete"`10622 - `status: optional "in_progress" or "completed" or "incomplete"`

10583 10623 

10584 The status of the item. One of `in_progress`, `completed`, or10624 The status of the item. One of `in_progress`, `completed`, or

11352 Unique identifier for the MCP tool call approval request.11392 Unique identifier for the MCP tool call approval request.

11353 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.11393 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.

11354 11394 

11355 - `error: optional string or null`11395 - `error: optional BetaMcpToolCallError or null`

11356 11396 

11357 The error from the tool call, if any.11397 The error from the tool call, if any.

11358 11398 

11399 - `McpProtocolError object { code, message, type }`

11400 

11401 - `code: number`

11402 

11403 - `message: string`

11404 

11405 - `type: "mcp_protocol_error"`

11406 

11407 - `"mcp_protocol_error"`

11408 

11409 - `McpToolExecutionError object { content, type }`

11410 

11411 - `content: unknown`

11412 

11413 - `type: "mcp_tool_execution_error"`

11414 

11415 - `"mcp_tool_execution_error"`

11416 

11417 - `HTTPError object { code, message, type }`

11418 

11419 - `code: number`

11420 

11421 - `message: string`

11422 

11423 - `type: "http_error"`

11424 

11425 - `"http_error"`

11426 

11359 - `output: optional string or null`11427 - `output: optional string or null`

11360 11428 

11361 The output from the tool call.11429 The output from the tool call.

12641 for reasoning items returned by `POST /v1/responses` and WebSocket12709 for reasoning items returned by `POST /v1/responses` and WebSocket

12642 `response.create` requests.12710 `response.create` requests.

12643 12711 

12712 When streaming, use the completed reasoning item and its

12713 `encrypted_content` from the `response.output_item.done` event in

12714 subsequent requests. The `encrypted_content` in

12715 `response.output_item.added` may be incomplete. This is especially

12716 important when `store` is `false` or when using Zero Data Retention.

12717 

12644 - `status: optional "in_progress" or "completed" or "incomplete"`12718 - `status: optional "in_progress" or "completed" or "incomplete"`

12645 12719 

12646 The status of the item. One of `in_progress`, `completed`, or12720 The status of the item. One of `in_progress`, `completed`, or

15518 Unique identifier for the MCP tool call approval request.15592 Unique identifier for the MCP tool call approval request.

15519 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.15593 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.

15520 15594 

15521 - `error: optional string or null`15595 - `error: optional BetaMcpToolCallError or null`

15522 15596 

15523 The error from the tool call, if any.15597 The error from the tool call, if any.

15524 15598 

3906 for reasoning items returned by `POST /v1/responses` and WebSocket3906 for reasoning items returned by `POST /v1/responses` and WebSocket

3907 `response.create` requests.3907 `response.create` requests.

3908 3908 

3909 When streaming, use the completed reasoning item and its

3910 `encrypted_content` from the `response.output_item.done` event in

3911 subsequent requests. The `encrypted_content` in

3912 `response.output_item.added` may be incomplete. This is especially

3913 important when `store` is `false` or when using Zero Data Retention.

3914 

3909 - `status: optional "in_progress" or "completed" or "incomplete"`3915 - `status: optional "in_progress" or "completed" or "incomplete"`

3910 3916 

3911 The status of the item. One of `in_progress`, `completed`, or3917 The status of the item. One of `in_progress`, `completed`, or

4679 Unique identifier for the MCP tool call approval request.4685 Unique identifier for the MCP tool call approval request.

4680 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.4686 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.

4681 4687 

4682 - `error: optional string or null`4688 - `error: optional BetaMcpToolCallError or null`

4683 4689 

4684 The error from the tool call, if any.4690 The error from the tool call, if any.

4685 4691 

4692 - `McpProtocolError object { code, message, type }`

4693 

4694 - `code: number`

4695 

4696 - `message: string`

4697 

4698 - `type: "mcp_protocol_error"`

4699 

4700 - `"mcp_protocol_error"`

4701 

4702 - `McpToolExecutionError object { content, type }`

4703 

4704 - `content: unknown`

4705 

4706 - `type: "mcp_tool_execution_error"`

4707 

4708 - `"mcp_tool_execution_error"`

4709 

4710 - `HTTPError object { code, message, type }`

4711 

4712 - `code: number`

4713 

4714 - `message: string`

4715 

4716 - `type: "http_error"`

4717 

4718 - `"http_error"`

4719 

4686 - `output: optional string or null`4720 - `output: optional string or null`

4687 4721 

4688 The output from the tool call.4722 The output from the tool call.

5968 for reasoning items returned by `POST /v1/responses` and WebSocket6002 for reasoning items returned by `POST /v1/responses` and WebSocket

5969 `response.create` requests.6003 `response.create` requests.

5970 6004 

6005 When streaming, use the completed reasoning item and its

6006 `encrypted_content` from the `response.output_item.done` event in

6007 subsequent requests. The `encrypted_content` in

6008 `response.output_item.added` may be incomplete. This is especially

6009 important when `store` is `false` or when using Zero Data Retention.

6010 

5971 - `status: optional "in_progress" or "completed" or "incomplete"`6011 - `status: optional "in_progress" or "completed" or "incomplete"`

5972 6012 

5973 The status of the item. One of `in_progress`, `completed`, or6013 The status of the item. One of `in_progress`, `completed`, or

8845 Unique identifier for the MCP tool call approval request.8885 Unique identifier for the MCP tool call approval request.

8846 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.8886 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.

8847 8887 

8848 - `error: optional string or null`8888 - `error: optional BetaMcpToolCallError or null`

8849 8889 

8850 The error from the tool call, if any.8890 The error from the tool call, if any.

8851 8891 

3719 for reasoning items returned by `POST /v1/responses` and WebSocket3719 for reasoning items returned by `POST /v1/responses` and WebSocket

3720 `response.create` requests.3720 `response.create` requests.

3721 3721 

3722 When streaming, use the completed reasoning item and its

3723 `encrypted_content` from the `response.output_item.done` event in

3724 subsequent requests. The `encrypted_content` in

3725 `response.output_item.added` may be incomplete. This is especially

3726 important when `store` is `false` or when using Zero Data Retention.

3727 

3722 - `status: optional "in_progress" or "completed" or "incomplete"`3728 - `status: optional "in_progress" or "completed" or "incomplete"`

3723 3729 

3724 The status of the item. One of `in_progress`, `completed`, or3730 The status of the item. One of `in_progress`, `completed`, or

4584 Unique identifier for the MCP tool call approval request.4590 Unique identifier for the MCP tool call approval request.

4585 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.4591 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.

4586 4592 

4587 - `error: optional string or null`4593 - `error: optional BetaMcpToolCallError or null`

4588 4594 

4589 The error from the tool call, if any.4595 The error from the tool call, if any.

4590 4596 

4597 - `McpProtocolError object { code, message, type }`

4598 

4599 - `code: number`

4600 

4601 - `message: string`

4602 

4603 - `type: "mcp_protocol_error"`

4604 

4605 - `"mcp_protocol_error"`

4606 

4607 - `McpToolExecutionError object { content, type }`

4608 

4609 - `content: unknown`

4610 

4611 - `type: "mcp_tool_execution_error"`

4612 

4613 - `"mcp_tool_execution_error"`

4614 

4615 - `HTTPError object { code, message, type }`

4616 

4617 - `code: number`

4618 

4619 - `message: string`

4620 

4621 - `type: "http_error"`

4622 

4623 - `"http_error"`

4624 

4591 - `output: optional string or null`4625 - `output: optional string or null`

4592 4626 

4593 The output from the tool call.4627 The output from the tool call.

3795 for reasoning items returned by `POST /v1/responses` and WebSocket3795 for reasoning items returned by `POST /v1/responses` and WebSocket

3796 `response.create` requests.3796 `response.create` requests.

3797 3797 

3798 When streaming, use the completed reasoning item and its

3799 `encrypted_content` from the `response.output_item.done` event in

3800 subsequent requests. The `encrypted_content` in

3801 `response.output_item.added` may be incomplete. This is especially

3802 important when `store` is `false` or when using Zero Data Retention.

3803 

3798 - `status: optional "in_progress" or "completed" or "incomplete"`3804 - `status: optional "in_progress" or "completed" or "incomplete"`

3799 3805 

3800 The status of the item. One of `in_progress`, `completed`, or3806 The status of the item. One of `in_progress`, `completed`, or

4568 Unique identifier for the MCP tool call approval request.4574 Unique identifier for the MCP tool call approval request.

4569 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.4575 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.

4570 4576 

4571 - `error: optional string or null`4577 - `error: optional BetaMcpToolCallError or null`

4572 4578 

4573 The error from the tool call, if any.4579 The error from the tool call, if any.

4574 4580 

4581 - `McpProtocolError object { code, message, type }`

4582 

4583 - `code: number`

4584 

4585 - `message: string`

4586 

4587 - `type: "mcp_protocol_error"`

4588 

4589 - `"mcp_protocol_error"`

4590 

4591 - `McpToolExecutionError object { content, type }`

4592 

4593 - `content: unknown`

4594 

4595 - `type: "mcp_tool_execution_error"`

4596 

4597 - `"mcp_tool_execution_error"`

4598 

4599 - `HTTPError object { code, message, type }`

4600 

4601 - `code: number`

4602 

4603 - `message: string`

4604 

4605 - `type: "http_error"`

4606 

4607 - `"http_error"`

4608 

4575 - `output: optional string or null`4609 - `output: optional string or null`

4576 4610 

4577 The output from the tool call.4611 The output from the tool call.

1394 1394 

1395### Returns1395### Returns

1396 1396 

1397- `ChatCompletion object { id, choices, created, 6 more }`1397- `ChatCompletion object { id, choices, created, 7 more }`

1398 1398 

1399 Represents a chat completion response returned by model, based on the provided input.1399 Represents a chat completion response returned by model, based on the provided input.

1400 1400 

1639 1639 

1640 - `"chat.completion"`1640 - `"chat.completion"`

1641 1641 

1642 - `metadata: optional Metadata or null`

1643 

1644 Set of 16 key-value pairs that can be attached to an object. This can be

1645 useful for storing additional information about the object in a structured

1646 format, and querying for objects via API or the dashboard.

1647 

1648 Keys are strings with a maximum length of 64 characters. Values are strings

1649 with a maximum length of 512 characters.

1650 

1642 - `moderation: optional object { input, output } or null`1651 - `moderation: optional object { input, output } or null`

1643 1652 

1644 Moderation results for the request input and generated output, if moderated1653 Moderation results for the request input and generated output, if moderated

1830 1839 

1831 Total number of tokens used in the request (prompt + completion).1840 Total number of tokens used in the request (prompt + completion).

1832 1841 

1833 - `completion_tokens_details: optional object { accepted_prediction_tokens, audio_tokens, reasoning_tokens, rejected_prediction_tokens }`1842 - `completion_tokens_details: optional object { accepted_prediction_tokens, audio_tokens, reasoning_tokens, 2 more }`

1834 1843 

1835 Breakdown of tokens used in a completion.1844 Breakdown of tokens used in a completion.

1836 1845 

1855 completion tokens for purposes of billing, output, and context window1864 completion tokens for purposes of billing, output, and context window

1856 limits.1865 limits.

1857 1866 

1858 - `prompt_tokens_details: optional object { audio_tokens, cache_write_tokens, cached_tokens }`1867 - `text_tokens: optional number`

1868 

1869 Text output tokens generated by the model.

1870 

1871 - `prompt_tokens_details: optional object { audio_tokens, cache_write_tokens, cached_tokens, 2 more }`

1859 1872 

1860 Breakdown of tokens used in the prompt.1873 Breakdown of tokens used in the prompt.

1861 1874 

1871 1884 

1872 Cached tokens present in the prompt.1885 Cached tokens present in the prompt.

1873 1886 

1887 - `image_tokens: optional number`

1888 

1889 Image input tokens present in the prompt.

1890 

1891 - `text_tokens: optional number`

1892 

1893 Text input tokens present in the prompt.

1894 

1874### Example1895### Example

1875 1896 

1876```http1897```http

1982 "created": 0,2003 "created": 0,

1983 "model": "model",2004 "model": "model",

1984 "object": "chat.completion",2005 "object": "chat.completion",

2006 "metadata": {

2007 "foo": "string"

2008 },

1985 "moderation": {2009 "moderation": {

1986 "input": {2010 "input": {

1987 "model": "model",2011 "model": "model",

2038 "accepted_prediction_tokens": 0,2062 "accepted_prediction_tokens": 0,

2039 "audio_tokens": 0,2063 "audio_tokens": 0,

2040 "reasoning_tokens": 0,2064 "reasoning_tokens": 0,

2041 "rejected_prediction_tokens": 02065 "rejected_prediction_tokens": 0,

2066 "text_tokens": 0

2042 },2067 },

2043 "prompt_tokens_details": {2068 "prompt_tokens_details": {

2044 "audio_tokens": 0,2069 "audio_tokens": 0,

2045 "cache_write_tokens": 0,2070 "cache_write_tokens": 0,

2046 "cached_tokens": 02071 "cached_tokens": 0,

2072 "image_tokens": 0,

2073 "text_tokens": 0

2047 }2074 }

2048 }2075 }

2049}2076}

280 280 

281 - `"chat.completion"`281 - `"chat.completion"`

282 282 

283 - `metadata: optional Metadata or null`

284 

285 Set of 16 key-value pairs that can be attached to an object. This can be

286 useful for storing additional information about the object in a structured

287 format, and querying for objects via API or the dashboard.

288 

289 Keys are strings with a maximum length of 64 characters. Values are strings

290 with a maximum length of 512 characters.

291 

283 - `moderation: optional object { input, output } or null`292 - `moderation: optional object { input, output } or null`

284 293 

285 Moderation results for the request input and generated output, if moderated294 Moderation results for the request input and generated output, if moderated

471 480 

472 Total number of tokens used in the request (prompt + completion).481 Total number of tokens used in the request (prompt + completion).

473 482 

474 - `completion_tokens_details: optional object { accepted_prediction_tokens, audio_tokens, reasoning_tokens, rejected_prediction_tokens }`483 - `completion_tokens_details: optional object { accepted_prediction_tokens, audio_tokens, reasoning_tokens, 2 more }`

475 484 

476 Breakdown of tokens used in a completion.485 Breakdown of tokens used in a completion.

477 486 

496 completion tokens for purposes of billing, output, and context window505 completion tokens for purposes of billing, output, and context window

497 limits.506 limits.

498 507 

499 - `prompt_tokens_details: optional object { audio_tokens, cache_write_tokens, cached_tokens }`508 - `text_tokens: optional number`

509 

510 Text output tokens generated by the model.

511 

512 - `prompt_tokens_details: optional object { audio_tokens, cache_write_tokens, cached_tokens, 2 more }`

500 513 

501 Breakdown of tokens used in the prompt.514 Breakdown of tokens used in the prompt.

502 515 

512 525 

513 Cached tokens present in the prompt.526 Cached tokens present in the prompt.

514 527 

528 - `image_tokens: optional number`

529 

530 Image input tokens present in the prompt.

531 

532 - `text_tokens: optional number`

533 

534 Text input tokens present in the prompt.

535 

515- `first_id: string`536- `first_id: string`

516 537 

517 The identifier of the first chat completion in the data array.538 The identifier of the first chat completion in the data array.

627 "created": 0,648 "created": 0,

628 "model": "model",649 "model": "model",

629 "object": "chat.completion",650 "object": "chat.completion",

651 "metadata": {

652 "foo": "string"

653 },

630 "moderation": {654 "moderation": {

631 "input": {655 "input": {

632 "model": "model",656 "model": "model",

683 "accepted_prediction_tokens": 0,707 "accepted_prediction_tokens": 0,

684 "audio_tokens": 0,708 "audio_tokens": 0,

685 "reasoning_tokens": 0,709 "reasoning_tokens": 0,

686 "rejected_prediction_tokens": 0710 "rejected_prediction_tokens": 0,

711 "text_tokens": 0

687 },712 },

688 "prompt_tokens_details": {713 "prompt_tokens_details": {

689 "audio_tokens": 0,714 "audio_tokens": 0,

690 "cache_write_tokens": 0,715 "cache_write_tokens": 0,

691 "cached_tokens": 0716 "cached_tokens": 0,

717 "image_tokens": 0,

718 "text_tokens": 0

692 }719 }

693 }720 }

694 }721 }

11 11 

12### Returns12### Returns

13 13 

14- `ChatCompletion object { id, choices, created, 6 more }`14- `ChatCompletion object { id, choices, created, 7 more }`

15 15 

16 Represents a chat completion response returned by model, based on the provided input.16 Represents a chat completion response returned by model, based on the provided input.

17 17 

256 256 

257 - `"chat.completion"`257 - `"chat.completion"`

258 258 

259 - `metadata: optional Metadata or null`

260 

261 Set of 16 key-value pairs that can be attached to an object. This can be

262 useful for storing additional information about the object in a structured

263 format, and querying for objects via API or the dashboard.

264 

265 Keys are strings with a maximum length of 64 characters. Values are strings

266 with a maximum length of 512 characters.

267 

259 - `moderation: optional object { input, output } or null`268 - `moderation: optional object { input, output } or null`

260 269 

261 Moderation results for the request input and generated output, if moderated270 Moderation results for the request input and generated output, if moderated

447 456 

448 Total number of tokens used in the request (prompt + completion).457 Total number of tokens used in the request (prompt + completion).

449 458 

450 - `completion_tokens_details: optional object { accepted_prediction_tokens, audio_tokens, reasoning_tokens, rejected_prediction_tokens }`459 - `completion_tokens_details: optional object { accepted_prediction_tokens, audio_tokens, reasoning_tokens, 2 more }`

451 460 

452 Breakdown of tokens used in a completion.461 Breakdown of tokens used in a completion.

453 462 

472 completion tokens for purposes of billing, output, and context window481 completion tokens for purposes of billing, output, and context window

473 limits.482 limits.

474 483 

475 - `prompt_tokens_details: optional object { audio_tokens, cache_write_tokens, cached_tokens }`484 - `text_tokens: optional number`

485 

486 Text output tokens generated by the model.

487 

488 - `prompt_tokens_details: optional object { audio_tokens, cache_write_tokens, cached_tokens, 2 more }`

476 489 

477 Breakdown of tokens used in the prompt.490 Breakdown of tokens used in the prompt.

478 491 

488 501 

489 Cached tokens present in the prompt.502 Cached tokens present in the prompt.

490 503 

504 - `image_tokens: optional number`

505 

506 Image input tokens present in the prompt.

507 

508 - `text_tokens: optional number`

509 

510 Text input tokens present in the prompt.

511 

491### Example512### Example

492 513 

493```http514```http

583 "created": 0,604 "created": 0,

584 "model": "model",605 "model": "model",

585 "object": "chat.completion",606 "object": "chat.completion",

607 "metadata": {

608 "foo": "string"

609 },

586 "moderation": {610 "moderation": {

587 "input": {611 "input": {

588 "model": "model",612 "model": "model",

639 "accepted_prediction_tokens": 0,663 "accepted_prediction_tokens": 0,

640 "audio_tokens": 0,664 "audio_tokens": 0,

641 "reasoning_tokens": 0,665 "reasoning_tokens": 0,

642 "rejected_prediction_tokens": 0666 "rejected_prediction_tokens": 0,

667 "text_tokens": 0

643 },668 },

644 "prompt_tokens_details": {669 "prompt_tokens_details": {

645 "audio_tokens": 0,670 "audio_tokens": 0,

646 "cache_write_tokens": 0,671 "cache_write_tokens": 0,

647 "cached_tokens": 0672 "cached_tokens": 0,

673 "image_tokens": 0,

674 "text_tokens": 0

648 }675 }

649 }676 }

650}677}

23 23 

24### Returns24### Returns

25 25 

26- `ChatCompletion object { id, choices, created, 6 more }`26- `ChatCompletion object { id, choices, created, 7 more }`

27 27 

28 Represents a chat completion response returned by model, based on the provided input.28 Represents a chat completion response returned by model, based on the provided input.

29 29 

268 268 

269 - `"chat.completion"`269 - `"chat.completion"`

270 270 

271 - `metadata: optional Metadata or null`

272 

273 Set of 16 key-value pairs that can be attached to an object. This can be

274 useful for storing additional information about the object in a structured

275 format, and querying for objects via API or the dashboard.

276 

277 Keys are strings with a maximum length of 64 characters. Values are strings

278 with a maximum length of 512 characters.

279 

271 - `moderation: optional object { input, output } or null`280 - `moderation: optional object { input, output } or null`

272 281 

273 Moderation results for the request input and generated output, if moderated282 Moderation results for the request input and generated output, if moderated

459 468 

460 Total number of tokens used in the request (prompt + completion).469 Total number of tokens used in the request (prompt + completion).

461 470 

462 - `completion_tokens_details: optional object { accepted_prediction_tokens, audio_tokens, reasoning_tokens, rejected_prediction_tokens }`471 - `completion_tokens_details: optional object { accepted_prediction_tokens, audio_tokens, reasoning_tokens, 2 more }`

463 472 

464 Breakdown of tokens used in a completion.473 Breakdown of tokens used in a completion.

465 474 

484 completion tokens for purposes of billing, output, and context window493 completion tokens for purposes of billing, output, and context window

485 limits.494 limits.

486 495 

487 - `prompt_tokens_details: optional object { audio_tokens, cache_write_tokens, cached_tokens }`496 - `text_tokens: optional number`

497 

498 Text output tokens generated by the model.

499 

500 - `prompt_tokens_details: optional object { audio_tokens, cache_write_tokens, cached_tokens, 2 more }`

488 501 

489 Breakdown of tokens used in the prompt.502 Breakdown of tokens used in the prompt.

490 503 

500 513 

501 Cached tokens present in the prompt.514 Cached tokens present in the prompt.

502 515 

516 - `image_tokens: optional number`

517 

518 Image input tokens present in the prompt.

519 

520 - `text_tokens: optional number`

521 

522 Text input tokens present in the prompt.

523 

503### Example524### Example

504 525 

505```http526```http

601 "created": 0,622 "created": 0,

602 "model": "model",623 "model": "model",

603 "object": "chat.completion",624 "object": "chat.completion",

625 "metadata": {

626 "foo": "string"

627 },

604 "moderation": {628 "moderation": {

605 "input": {629 "input": {

606 "model": "model",630 "model": "model",

657 "accepted_prediction_tokens": 0,681 "accepted_prediction_tokens": 0,

658 "audio_tokens": 0,682 "audio_tokens": 0,

659 "reasoning_tokens": 0,683 "reasoning_tokens": 0,

660 "rejected_prediction_tokens": 0684 "rejected_prediction_tokens": 0,

685 "text_tokens": 0

661 },686 },

662 "prompt_tokens_details": {687 "prompt_tokens_details": {

663 "audio_tokens": 0,688 "audio_tokens": 0,

664 "cache_write_tokens": 0,689 "cache_write_tokens": 0,

665 "cached_tokens": 0690 "cached_tokens": 0,

691 "image_tokens": 0,

692 "text_tokens": 0

666 }693 }

667 }694 }

668}695}

624 },624 },

625 {625 {

626 "ident": "rejected_prediction_tokens"626 "ident": "rejected_prediction_tokens"

627 },

628 {

629 "ident": "text_tokens"

627 }630 }

628 ]631 ]

629 },632 },

635 "(resource) completions > (model) completion_usage > (schema) > (property) completion_tokens_details > (property) accepted_prediction_tokens",638 "(resource) completions > (model) completion_usage > (schema) > (property) completion_tokens_details > (property) accepted_prediction_tokens",

636 "(resource) completions > (model) completion_usage > (schema) > (property) completion_tokens_details > (property) audio_tokens",639 "(resource) completions > (model) completion_usage > (schema) > (property) completion_tokens_details > (property) audio_tokens",

637 "(resource) completions > (model) completion_usage > (schema) > (property) completion_tokens_details > (property) reasoning_tokens",640 "(resource) completions > (model) completion_usage > (schema) > (property) completion_tokens_details > (property) reasoning_tokens",

638 "(resource) completions > (model) completion_usage > (schema) > (property) completion_tokens_details > (property) rejected_prediction_tokens"641 "(resource) completions > (model) completion_usage > (schema) > (property) completion_tokens_details > (property) rejected_prediction_tokens",

642 "(resource) completions > (model) completion_usage > (schema) > (property) completion_tokens_details > (property) text_tokens"

639 ]643 ]

640 },644 },

641 "(resource) completions > (model) completion_usage > (schema) > (property) prompt_tokens_details": {645 "(resource) completions > (model) completion_usage > (schema) > (property) prompt_tokens_details": {

655 },659 },

656 {660 {

657 "ident": "cached_tokens"661 "ident": "cached_tokens"

662 },

663 {

664 "ident": "image_tokens"

665 },

666 {

667 "ident": "text_tokens"

658 }668 }

659 ]669 ]

660 },670 },

665 "children": [675 "children": [

666 "(resource) completions > (model) completion_usage > (schema) > (property) prompt_tokens_details > (property) audio_tokens",676 "(resource) completions > (model) completion_usage > (schema) > (property) prompt_tokens_details > (property) audio_tokens",

667 "(resource) completions > (model) completion_usage > (schema) > (property) prompt_tokens_details > (property) cache_write_tokens",677 "(resource) completions > (model) completion_usage > (schema) > (property) prompt_tokens_details > (property) cache_write_tokens",

668 "(resource) completions > (model) completion_usage > (schema) > (property) prompt_tokens_details > (property) cached_tokens"678 "(resource) completions > (model) completion_usage > (schema) > (property) prompt_tokens_details > (property) cached_tokens",

679 "(resource) completions > (model) completion_usage > (schema) > (property) prompt_tokens_details > (property) image_tokens",

680 "(resource) completions > (model) completion_usage > (schema) > (property) prompt_tokens_details > (property) text_tokens"

669 ]681 ]

670 },682 },

671 "(resource) completions > (model) completion_usage > (schema)": {683 "(resource) completions > (model) completion_usage > (schema)": {

1088 "schemaType": "integer",1100 "schemaType": "integer",

1089 "children": []1101 "children": []

1090 },1102 },

1103 "(resource) completions > (model) completion_usage > (schema) > (property) completion_tokens_details > (property) text_tokens": {

1104 "kind": "HttpDeclProperty",

1105 "oasRef": "#/components/schemas/CompletionUsage/properties/completion_tokens_details/properties/text_tokens",

1106 "deprecated": false,

1107 "key": "text_tokens",

1108 "docstring": "Text output tokens generated by the model.",

1109 "type": {

1110 "kind": "HttpTypeNumber"

1111 },

1112 "optional": true,

1113 "nullable": false,

1114 "schemaType": "integer",

1115 "children": []

1116 },

1091 "(resource) completions > (model) completion_usage > (schema) > (property) prompt_tokens_details > (property) audio_tokens": {1117 "(resource) completions > (model) completion_usage > (schema) > (property) prompt_tokens_details > (property) audio_tokens": {

1092 "kind": "HttpDeclProperty",1118 "kind": "HttpDeclProperty",

1093 "oasRef": "#/components/schemas/CompletionUsage/properties/prompt_tokens_details/properties/audio_tokens",1119 "oasRef": "#/components/schemas/CompletionUsage/properties/prompt_tokens_details/properties/audio_tokens",

1133 "schemaType": "integer",1159 "schemaType": "integer",

1134 "children": []1160 "children": []

1135 },1161 },

1162 "(resource) completions > (model) completion_usage > (schema) > (property) prompt_tokens_details > (property) image_tokens": {

1163 "kind": "HttpDeclProperty",

1164 "oasRef": "#/components/schemas/CompletionUsage/properties/prompt_tokens_details/properties/image_tokens",

1165 "deprecated": false,

1166 "key": "image_tokens",

1167 "docstring": "Image input tokens present in the prompt.",

1168 "type": {

1169 "kind": "HttpTypeNumber"

1170 },

1171 "optional": true,

1172 "nullable": false,

1173 "schemaType": "integer",

1174 "children": []

1175 },

1176 "(resource) completions > (model) completion_usage > (schema) > (property) prompt_tokens_details > (property) text_tokens": {

1177 "kind": "HttpDeclProperty",

1178 "oasRef": "#/components/schemas/CompletionUsage/properties/prompt_tokens_details/properties/text_tokens",

1179 "deprecated": false,

1180 "key": "text_tokens",

1181 "docstring": "Text input tokens present in the prompt.",

1182 "type": {

1183 "kind": "HttpTypeNumber"

1184 },

1185 "optional": true,

1186 "nullable": false,

1187 "schemaType": "integer",

1188 "children": []

1189 },

1136 "(resource) chat.completions > (model) chat_completion_chunk > (schema) > (property) choices > (items) > (property) delta > (property) function_call > (property) arguments": {1190 "(resource) chat.completions > (model) chat_completion_chunk > (schema) > (property) choices > (items) > (property) delta > (property) function_call > (property) arguments": {

1137 "kind": "HttpDeclProperty",1191 "kind": "HttpDeclProperty",

1138 "oasRef": "#/components/schemas/ChatCompletionStreamResponseDelta/properties/function_call/properties/arguments",1192 "oasRef": "#/components/schemas/ChatCompletionStreamResponseDelta/properties/function_call/properties/arguments",

232 232 

233 Total number of tokens used in the request (prompt + completion).233 Total number of tokens used in the request (prompt + completion).

234 234 

235 - `completion_tokens_details: optional object { accepted_prediction_tokens, audio_tokens, reasoning_tokens, rejected_prediction_tokens }`235 - `completion_tokens_details: optional object { accepted_prediction_tokens, audio_tokens, reasoning_tokens, 2 more }`

236 236 

237 Breakdown of tokens used in a completion.237 Breakdown of tokens used in a completion.

238 238 

257 completion tokens for purposes of billing, output, and context window257 completion tokens for purposes of billing, output, and context window

258 limits.258 limits.

259 259 

260 - `prompt_tokens_details: optional object { audio_tokens, cache_write_tokens, cached_tokens }`260 - `text_tokens: optional number`

261 

262 Text output tokens generated by the model.

263 

264 - `prompt_tokens_details: optional object { audio_tokens, cache_write_tokens, cached_tokens, 2 more }`

261 265 

262 Breakdown of tokens used in the prompt.266 Breakdown of tokens used in the prompt.

263 267 

273 277 

274 Cached tokens present in the prompt.278 Cached tokens present in the prompt.

275 279 

280 - `image_tokens: optional number`

281 

282 Image input tokens present in the prompt.

283 

284 - `text_tokens: optional number`

285 

286 Text input tokens present in the prompt.

287 

276### Example288### Example

277 289 

278```http290```http

331 "accepted_prediction_tokens": 0,343 "accepted_prediction_tokens": 0,

332 "audio_tokens": 0,344 "audio_tokens": 0,

333 "reasoning_tokens": 0,345 "reasoning_tokens": 0,

334 "rejected_prediction_tokens": 0346 "rejected_prediction_tokens": 0,

347 "text_tokens": 0

335 },348 },

336 "prompt_tokens_details": {349 "prompt_tokens_details": {

337 "audio_tokens": 0,350 "audio_tokens": 0,

338 "cache_write_tokens": 0,351 "cache_write_tokens": 0,

339 "cached_tokens": 0352 "cached_tokens": 0,

353 "image_tokens": 0,

354 "text_tokens": 0

340 }355 }

341 }356 }

342}357}

3175 for reasoning items returned by `POST /v1/responses` and WebSocket3175 for reasoning items returned by `POST /v1/responses` and WebSocket

3176 `response.create` requests.3176 `response.create` requests.

3177 3177 

3178 When streaming, use the completed reasoning item and its

3179 `encrypted_content` from the `response.output_item.done` event in

3180 subsequent requests. The `encrypted_content` in

3181 `response.output_item.added` may be incomplete. This is especially

3182 important when `store` is `false` or when using Zero Data Retention.

3183 

3178 - `status: optional "in_progress" or "completed" or "incomplete"`3184 - `status: optional "in_progress" or "completed" or "incomplete"`

3179 3185 

3180 The status of the item. One of `in_progress`, `completed`, or3186 The status of the item. One of `in_progress`, `completed`, or

3844 Unique identifier for the MCP tool call approval request.3850 Unique identifier for the MCP tool call approval request.

3845 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.3851 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.

3846 3852 

3847 - `error: optional string or null`3853 - `error: optional McpToolCallError or null`

3848 3854 

3849 The error from the tool call, if any.3855 The error from the tool call, if any.

3850 3856 

3857 - `McpProtocolError object { code, message, type }`

3858 

3859 - `code: number`

3860 

3861 - `message: string`

3862 

3863 - `type: "mcp_protocol_error"`

3864 

3865 - `"mcp_protocol_error"`

3866 

3867 - `McpToolExecutionError object { content, type }`

3868 

3869 - `content: unknown`

3870 

3871 - `type: "mcp_tool_execution_error"`

3872 

3873 - `"mcp_tool_execution_error"`

3874 

3875 - `HTTPError object { code, message, type }`

3876 

3877 - `code: number`

3878 

3879 - `message: string`

3880 

3881 - `type: "http_error"`

3882 

3883 - `"http_error"`

3884 

3851 - `output: optional string or null`3885 - `output: optional string or null`

3852 3886 

3853 The output from the tool call.3887 The output from the tool call.

3202 for reasoning items returned by `POST /v1/responses` and WebSocket3202 for reasoning items returned by `POST /v1/responses` and WebSocket

3203 `response.create` requests.3203 `response.create` requests.

3204 3204 

3205 When streaming, use the completed reasoning item and its

3206 `encrypted_content` from the `response.output_item.done` event in

3207 subsequent requests. The `encrypted_content` in

3208 `response.output_item.added` may be incomplete. This is especially

3209 important when `store` is `false` or when using Zero Data Retention.

3210 

3205 - `status: optional "in_progress" or "completed" or "incomplete"`3211 - `status: optional "in_progress" or "completed" or "incomplete"`

3206 3212 

3207 The status of the item. One of `in_progress`, `completed`, or3213 The status of the item. One of `in_progress`, `completed`, or

3871 Unique identifier for the MCP tool call approval request.3877 Unique identifier for the MCP tool call approval request.

3872 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.3878 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.

3873 3879 

3874 - `error: optional string or null`3880 - `error: optional McpToolCallError or null`

3875 3881 

3876 The error from the tool call, if any.3882 The error from the tool call, if any.

3877 3883 

3884 - `McpProtocolError object { code, message, type }`

3885 

3886 - `code: number`

3887 

3888 - `message: string`

3889 

3890 - `type: "mcp_protocol_error"`

3891 

3892 - `"mcp_protocol_error"`

3893 

3894 - `McpToolExecutionError object { content, type }`

3895 

3896 - `content: unknown`

3897 

3898 - `type: "mcp_tool_execution_error"`

3899 

3900 - `"mcp_tool_execution_error"`

3901 

3902 - `HTTPError object { code, message, type }`

3903 

3904 - `code: number`

3905 

3906 - `message: string`

3907 

3908 - `type: "http_error"`

3909 

3910 - `"http_error"`

3911 

3878 - `output: optional string or null`3912 - `output: optional string or null`

3879 3913 

3880 The output from the tool call.3914 The output from the tool call.

7224 for reasoning items returned by `POST /v1/responses` and WebSocket7258 for reasoning items returned by `POST /v1/responses` and WebSocket

7225 `response.create` requests.7259 `response.create` requests.

7226 7260 

7261 When streaming, use the completed reasoning item and its

7262 `encrypted_content` from the `response.output_item.done` event in

7263 subsequent requests. The `encrypted_content` in

7264 `response.output_item.added` may be incomplete. This is especially

7265 important when `store` is `false` or when using Zero Data Retention.

7266 

7227 - `status: optional "in_progress" or "completed" or "incomplete"`7267 - `status: optional "in_progress" or "completed" or "incomplete"`

7228 7268 

7229 The status of the item. One of `in_progress`, `completed`, or7269 The status of the item. One of `in_progress`, `completed`, or

7939 Unique identifier for the MCP tool call approval request.7979 Unique identifier for the MCP tool call approval request.

7940 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.7980 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.

7941 7981 

7942 - `error: optional string or null`7982 - `error: optional McpToolCallError or null`

7943 7983 

7944 The error from the tool call, if any.7984 The error from the tool call, if any.

7945 7985 

7986 - `McpProtocolError object { code, message, type }`

7987 

7988 - `code: number`

7989 

7990 - `message: string`

7991 

7992 - `type: "mcp_protocol_error"`

7993 

7994 - `"mcp_protocol_error"`

7995 

7996 - `McpToolExecutionError object { content, type }`

7997 

7998 - `content: unknown`

7999 

8000 - `type: "mcp_tool_execution_error"`

8001 

8002 - `"mcp_tool_execution_error"`

8003 

8004 - `HTTPError object { code, message, type }`

8005 

8006 - `code: number`

8007 

8008 - `message: string`

8009 

8010 - `type: "http_error"`

8011 

8012 - `"http_error"`

8013 

7946 - `output: optional string or null`8014 - `output: optional string or null`

7947 8015 

7948 The output from the tool call.8016 The output from the tool call.

3196 for reasoning items returned by `POST /v1/responses` and WebSocket3196 for reasoning items returned by `POST /v1/responses` and WebSocket

3197 `response.create` requests.3197 `response.create` requests.

3198 3198 

3199 When streaming, use the completed reasoning item and its

3200 `encrypted_content` from the `response.output_item.done` event in

3201 subsequent requests. The `encrypted_content` in

3202 `response.output_item.added` may be incomplete. This is especially

3203 important when `store` is `false` or when using Zero Data Retention.

3204 

3199 - `status: optional "in_progress" or "completed" or "incomplete"`3205 - `status: optional "in_progress" or "completed" or "incomplete"`

3200 3206 

3201 The status of the item. One of `in_progress`, `completed`, or3207 The status of the item. One of `in_progress`, `completed`, or

3911 Unique identifier for the MCP tool call approval request.3917 Unique identifier for the MCP tool call approval request.

3912 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.3918 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.

3913 3919 

3914 - `error: optional string or null`3920 - `error: optional McpToolCallError or null`

3915 3921 

3916 The error from the tool call, if any.3922 The error from the tool call, if any.

3917 3923 

3924 - `McpProtocolError object { code, message, type }`

3925 

3926 - `code: number`

3927 

3928 - `message: string`

3929 

3930 - `type: "mcp_protocol_error"`

3931 

3932 - `"mcp_protocol_error"`

3933 

3934 - `McpToolExecutionError object { content, type }`

3935 

3936 - `content: unknown`

3937 

3938 - `type: "mcp_tool_execution_error"`

3939 

3940 - `"mcp_tool_execution_error"`

3941 

3942 - `HTTPError object { code, message, type }`

3943 

3944 - `code: number`

3945 

3946 - `message: string`

3947 

3948 - `type: "http_error"`

3949 

3950 - `"http_error"`

3951 

3918 - `output: optional string or null`3952 - `output: optional string or null`

3919 3953 

3920 The output from the tool call.3954 The output from the tool call.

3167 for reasoning items returned by `POST /v1/responses` and WebSocket3167 for reasoning items returned by `POST /v1/responses` and WebSocket

3168 `response.create` requests.3168 `response.create` requests.

3169 3169 

3170 When streaming, use the completed reasoning item and its

3171 `encrypted_content` from the `response.output_item.done` event in

3172 subsequent requests. The `encrypted_content` in

3173 `response.output_item.added` may be incomplete. This is especially

3174 important when `store` is `false` or when using Zero Data Retention.

3175 

3170 - `status: optional "in_progress" or "completed" or "incomplete"`3176 - `status: optional "in_progress" or "completed" or "incomplete"`

3171 3177 

3172 The status of the item. One of `in_progress`, `completed`, or3178 The status of the item. One of `in_progress`, `completed`, or

3882 Unique identifier for the MCP tool call approval request.3888 Unique identifier for the MCP tool call approval request.

3883 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.3889 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.

3884 3890 

3885 - `error: optional string or null`3891 - `error: optional McpToolCallError or null`

3886 3892 

3887 The error from the tool call, if any.3893 The error from the tool call, if any.

3888 3894 

3895 - `McpProtocolError object { code, message, type }`

3896 

3897 - `code: number`

3898 

3899 - `message: string`

3900 

3901 - `type: "mcp_protocol_error"`

3902 

3903 - `"mcp_protocol_error"`

3904 

3905 - `McpToolExecutionError object { content, type }`

3906 

3907 - `content: unknown`

3908 

3909 - `type: "mcp_tool_execution_error"`

3910 

3911 - `"mcp_tool_execution_error"`

3912 

3913 - `HTTPError object { code, message, type }`

3914 

3915 - `code: number`

3916 

3917 - `message: string`

3918 

3919 - `type: "http_error"`

3920 

3921 - `"http_error"`

3922 

3889 - `output: optional string or null`3923 - `output: optional string or null`

3890 3924 

3891 The output from the tool call.3925 The output from the tool call.

26 26 

27 The organization that owns the model.27 The organization that owns the model.

28 28 

29 - `shutdown_date: optional string or null`

30 

31 The date when the model will shut down, or null if not announced.

32 

29- `object: "list"`33- `object: "list"`

30 34 

31 - `"list"`35 - `"list"`

46 "id": "id",50 "id": "id",

47 "created": 0,51 "created": 0,

48 "object": "model",52 "object": "model",

49 "owned_by": "owned_by"53 "owned_by": "owned_by",

54 "shutdown_date": "2019-12-27"

50 }55 }

51 ],56 ],

52 "object": "list"57 "object": "list"

70 "id": "model-id-0",75 "id": "model-id-0",

71 "object": "model",76 "object": "model",

72 "created": 1686935002,77 "created": 1686935002,

73 "owned_by": "organization-owner"78 "owned_by": "organization-owner",

79 "shutdown_date": null

74 },80 },

75 {81 {

76 "id": "model-id-1",82 "id": "model-id-1",

77 "object": "model",83 "object": "model",

78 "created": 1686935002,84 "created": 1686935002,

79 "owned_by": "organization-owner",85 "owned_by": "organization-owner",

86 "shutdown_date": null

80 },87 },

81 {88 {

82 "id": "model-id-2",89 "id": "model-id-2",

83 "object": "model",90 "object": "model",

84 "created": 1686935002,91 "created": 1686935002,

85 "owned_by": "openai"92 "owned_by": "openai",

93 "shutdown_date": "2026-10-23"

86 },94 },

87 ]95 ]

88}96}

10 10 

11### Returns11### Returns

12 12 

13- `Model object { id, created, object, owned_by }`13- `Model object { id, created, object, 2 more }`

14 14 

15 Describes an OpenAI model offering that can be used with the API.15 Describes an OpenAI model offering that can be used with the API.

16 16 

32 32 

33 The organization that owns the model.33 The organization that owns the model.

34 34 

35 - `shutdown_date: optional string or null`

36 

37 The date when the model will shut down, or null if not announced.

38 

35### Example39### Example

36 40 

37```http41```http

46 "id": "id",50 "id": "id",

47 "created": 0,51 "created": 0,

48 "object": "model",52 "object": "model",

49 "owned_by": "owned_by"53 "owned_by": "owned_by",

54 "shutdown_date": "2019-12-27"

50}55}

51```56```

52 57 

64 "id": "VAR_chat_model_id",69 "id": "VAR_chat_model_id",

65 "object": "model",70 "object": "model",

66 "created": 1686935002,71 "created": 1686935002,

67 "owned_by": "openai"72 "owned_by": "openai",

73 "shutdown_date": "2026-10-23"

68}74}

69```75```

3269 for reasoning items returned by `POST /v1/responses` and WebSocket3269 for reasoning items returned by `POST /v1/responses` and WebSocket

3270 `response.create` requests.3270 `response.create` requests.

3271 3271 

3272 When streaming, use the completed reasoning item and its

3273 `encrypted_content` from the `response.output_item.done` event in

3274 subsequent requests. The `encrypted_content` in

3275 `response.output_item.added` may be incomplete. This is especially

3276 important when `store` is `false` or when using Zero Data Retention.

3277 

3272 - `status: optional "in_progress" or "completed" or "incomplete"`3278 - `status: optional "in_progress" or "completed" or "incomplete"`

3273 3279 

3274 The status of the item. One of `in_progress`, `completed`, or3280 The status of the item. One of `in_progress`, `completed`, or

3938 Unique identifier for the MCP tool call approval request.3944 Unique identifier for the MCP tool call approval request.

3939 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.3945 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.

3940 3946 

3941 - `error: optional string or null`3947 - `error: optional McpToolCallError or null`

3942 3948 

3943 The error from the tool call, if any.3949 The error from the tool call, if any.

3944 3950 

3951 - `McpProtocolError object { code, message, type }`

3952 

3953 - `code: number`

3954 

3955 - `message: string`

3956 

3957 - `type: "mcp_protocol_error"`

3958 

3959 - `"mcp_protocol_error"`

3960 

3961 - `McpToolExecutionError object { content, type }`

3962 

3963 - `content: unknown`

3964 

3965 - `type: "mcp_tool_execution_error"`

3966 

3967 - `"mcp_tool_execution_error"`

3968 

3969 - `HTTPError object { code, message, type }`

3970 

3971 - `code: number`

3972 

3973 - `message: string`

3974 

3975 - `type: "http_error"`

3976 

3977 - `"http_error"`

3978 

3945 - `output: optional string or null`3979 - `output: optional string or null`

3946 3980 

3947 The output from the tool call.3981 The output from the tool call.

4872 for reasoning items returned by `POST /v1/responses` and WebSocket4906 for reasoning items returned by `POST /v1/responses` and WebSocket

4873 `response.create` requests.4907 `response.create` requests.

4874 4908 

4909 When streaming, use the completed reasoning item and its

4910 `encrypted_content` from the `response.output_item.done` event in

4911 subsequent requests. The `encrypted_content` in

4912 `response.output_item.added` may be incomplete. This is especially

4913 important when `store` is `false` or when using Zero Data Retention.

4914 

4875 - `status: optional "in_progress" or "completed" or "incomplete"`4915 - `status: optional "in_progress" or "completed" or "incomplete"`

4876 4916 

4877 The status of the item. One of `in_progress`, `completed`, or4917 The status of the item. One of `in_progress`, `completed`, or

7269 Unique identifier for the MCP tool call approval request.7309 Unique identifier for the MCP tool call approval request.

7270 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.7310 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.

7271 7311 

7272 - `error: optional string or null`7312 - `error: optional McpToolCallError or null`

7273 7313 

7274 The error from the tool call, if any.7314 The error from the tool call, if any.

7275 7315 

3399 for reasoning items returned by `POST /v1/responses` and WebSocket3399 for reasoning items returned by `POST /v1/responses` and WebSocket

3400 `response.create` requests.3400 `response.create` requests.

3401 3401 

3402 When streaming, use the completed reasoning item and its

3403 `encrypted_content` from the `response.output_item.done` event in

3404 subsequent requests. The `encrypted_content` in

3405 `response.output_item.added` may be incomplete. This is especially

3406 important when `store` is `false` or when using Zero Data Retention.

3407 

3402 - `status: optional "in_progress" or "completed" or "incomplete"`3408 - `status: optional "in_progress" or "completed" or "incomplete"`

3403 3409 

3404 The status of the item. One of `in_progress`, `completed`, or3410 The status of the item. One of `in_progress`, `completed`, or

4068 Unique identifier for the MCP tool call approval request.4074 Unique identifier for the MCP tool call approval request.

4069 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.4075 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.

4070 4076 

4071 - `error: optional string or null`4077 - `error: optional McpToolCallError or null`

4072 4078 

4073 The error from the tool call, if any.4079 The error from the tool call, if any.

4074 4080 

4081 - `McpProtocolError object { code, message, type }`

4082 

4083 - `code: number`

4084 

4085 - `message: string`

4086 

4087 - `type: "mcp_protocol_error"`

4088 

4089 - `"mcp_protocol_error"`

4090 

4091 - `McpToolExecutionError object { content, type }`

4092 

4093 - `content: unknown`

4094 

4095 - `type: "mcp_tool_execution_error"`

4096 

4097 - `"mcp_tool_execution_error"`

4098 

4099 - `HTTPError object { code, message, type }`

4100 

4101 - `code: number`

4102 

4103 - `message: string`

4104 

4105 - `type: "http_error"`

4106 

4107 - `"http_error"`

4108 

4075 - `output: optional string or null`4109 - `output: optional string or null`

4076 4110 

4077 The output from the tool call.4111 The output from the tool call.

7537 for reasoning items returned by `POST /v1/responses` and WebSocket7571 for reasoning items returned by `POST /v1/responses` and WebSocket

7538 `response.create` requests.7572 `response.create` requests.

7539 7573 

7574 When streaming, use the completed reasoning item and its

7575 `encrypted_content` from the `response.output_item.done` event in

7576 subsequent requests. The `encrypted_content` in

7577 `response.output_item.added` may be incomplete. This is especially

7578 important when `store` is `false` or when using Zero Data Retention.

7579 

7540 - `status: optional "in_progress" or "completed" or "incomplete"`7580 - `status: optional "in_progress" or "completed" or "incomplete"`

7541 7581 

7542 The status of the item. One of `in_progress`, `completed`, or7582 The status of the item. One of `in_progress`, `completed`, or

8200 Unique identifier for the MCP tool call approval request.8240 Unique identifier for the MCP tool call approval request.

8201 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.8241 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.

8202 8242 

8203 - `error: optional string or null`8243 - `error: optional McpToolCallError or null`

8204 8244 

8205 The error from the tool call, if any.8245 The error from the tool call, if any.

8206 8246 

8247 - `McpProtocolError object { code, message, type }`

8248 

8249 - `code: number`

8250 

8251 - `message: string`

8252 

8253 - `type: "mcp_protocol_error"`

8254 

8255 - `"mcp_protocol_error"`

8256 

8257 - `McpToolExecutionError object { content, type }`

8258 

8259 - `content: unknown`

8260 

8261 - `type: "mcp_tool_execution_error"`

8262 

8263 - `"mcp_tool_execution_error"`

8264 

8265 - `HTTPError object { code, message, type }`

8266 

8267 - `code: number`

8268 

8269 - `message: string`

8270 

8271 - `type: "http_error"`

8272 

8273 - `"http_error"`

8274 

8207 - `output: optional string or null`8275 - `output: optional string or null`

8208 8276 

8209 The output from the tool call.8277 The output from the tool call.

3261 for reasoning items returned by `POST /v1/responses` and WebSocket3261 for reasoning items returned by `POST /v1/responses` and WebSocket

3262 `response.create` requests.3262 `response.create` requests.

3263 3263 

3264 When streaming, use the completed reasoning item and its

3265 `encrypted_content` from the `response.output_item.done` event in

3266 subsequent requests. The `encrypted_content` in

3267 `response.output_item.added` may be incomplete. This is especially

3268 important when `store` is `false` or when using Zero Data Retention.

3269 

3264 - `status: optional "in_progress" or "completed" or "incomplete"`3270 - `status: optional "in_progress" or "completed" or "incomplete"`

3265 3271 

3266 The status of the item. One of `in_progress`, `completed`, or3272 The status of the item. One of `in_progress`, `completed`, or

3930 Unique identifier for the MCP tool call approval request.3936 Unique identifier for the MCP tool call approval request.

3931 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.3937 Include this value in a subsequent `mcp_approval_response` input to approve or reject the corresponding tool call.

3932 3938 

3933 - `error: optional string or null`3939 - `error: optional McpToolCallError or null`

3934 3940 

3935 The error from the tool call, if any.3941 The error from the tool call, if any.

3936 3942 

3943 - `McpProtocolError object { code, message, type }`

3944 

3945 - `code: number`

3946 

3947 - `message: string`

3948 

3949 - `type: "mcp_protocol_error"`

3950