feat(api): add o1 models (openai#49)

See https://platform.openai.com/docs/guides/reasoning for details.

Author: Stainless Bot

.stats.yml

@@ -1,2 +1,2 @@
 configured_endpoints: 68
-openapi_spec_url: https://storage.googleapis.com/stainless-sdk-openapi-specs/openai-85a85e0c08de456441431c0ae4e9c078cc8f9748c29430b9a9058340db6389ee.yml
+openapi_spec_url: https://storage.googleapis.com/stainless-sdk-openapi-specs/openai-501122aa32adaa2abb3d4487880ab9cdf2141addce2e6c3d1bd9bb6b44c318a8.yml

betaassistant.go

@@ -1860,7 +1860,8 @@ type FileSearchToolFileSearch struct {
 	// [file search tool documentation](https://platform.openai.com/docs/assistants/tools/file-search/customizing-file-search-settings)
 	// for more information.
 	MaxNumResults int64 `json:"max_num_results"`
-	// The ranking options for the file search.
+	// The ranking options for the file search. If not specified, the file search tool
+	// will use the `auto` ranker and a score_threshold of 0.
 	//
 	// See the
 	// [file search tool documentation](https://platform.openai.com/docs/assistants/tools/file-search/customizing-file-search-settings)
@@ -1886,26 +1887,27 @@ func (r fileSearchToolFileSearchJSON) RawJSON() string {
 	return r.raw
 }
 
-// The ranking options for the file search.
+// The ranking options for the file search. If not specified, the file search tool
+// will use the `auto` ranker and a score_threshold of 0.
 //
 // See the
 // [file search tool documentation](https://platform.openai.com/docs/assistants/tools/file-search/customizing-file-search-settings)
 // for more information.
 type FileSearchToolFileSearchRankingOptions struct {
+	// The score threshold for the file search. All values must be a floating point
+	// number between 0 and 1.
+	ScoreThreshold float64 `json:"score_threshold,required"`
 	// The ranker to use for the file search. If not specified will use the `auto`
 	// ranker.
 	Ranker FileSearchToolFileSearchRankingOptionsRanker `json:"ranker"`
-	// The score threshold for the file search. All values must be a floating point
-	// number between 0 and 1.
-	ScoreThreshold float64                                    `json:"score_threshold"`
-	JSON           fileSearchToolFileSearchRankingOptionsJSON `json:"-"`
+	JSON   fileSearchToolFileSearchRankingOptionsJSON   `json:"-"`
 }
 
 // fileSearchToolFileSearchRankingOptionsJSON contains the JSON metadata for the
 // struct [FileSearchToolFileSearchRankingOptions]
 type fileSearchToolFileSearchRankingOptionsJSON struct {
-	Ranker         apijson.Field
 	ScoreThreshold apijson.Field
+	Ranker         apijson.Field
 	raw            string
 	ExtraFields    map[string]apijson.Field
 }
@@ -1961,7 +1963,8 @@ type FileSearchToolFileSearchParam struct {
 	// [file search tool documentation](https://platform.openai.com/docs/assistants/tools/file-search/customizing-file-search-settings)
 	// for more information.
 	MaxNumResults param.Field[int64] `json:"max_num_results"`
-	// The ranking options for the file search.
+	// The ranking options for the file search. If not specified, the file search tool
+	// will use the `auto` ranker and a score_threshold of 0.
 	//
 	// See the
 	// [file search tool documentation](https://platform.openai.com/docs/assistants/tools/file-search/customizing-file-search-settings)
@@ -1973,18 +1976,19 @@ func (r FileSearchToolFileSearchParam) MarshalJSON() (data []byte, err error) {
 	return apijson.MarshalRoot(r)
 }
 
-// The ranking options for the file search.
+// The ranking options for the file search. If not specified, the file search tool
+// will use the `auto` ranker and a score_threshold of 0.
 //
 // See the
 // [file search tool documentation](https://platform.openai.com/docs/assistants/tools/file-search/customizing-file-search-settings)
 // for more information.
 type FileSearchToolFileSearchRankingOptionsParam struct {
+	// The score threshold for the file search. All values must be a floating point
+	// number between 0 and 1.
+	ScoreThreshold param.Field[float64] `json:"score_threshold,required"`
 	// The ranker to use for the file search. If not specified will use the `auto`
 	// ranker.
 	Ranker param.Field[FileSearchToolFileSearchRankingOptionsRanker] `json:"ranker"`
-	// The score threshold for the file search. All values must be a floating point
-	// number between 0 and 1.
-	ScoreThreshold param.Field[float64] `json:"score_threshold"`
 }
 
 func (r FileSearchToolFileSearchRankingOptionsParam) MarshalJSON() (data []byte, err error) {

chat.go

@@ -30,9 +30,13 @@ func NewChatService(opts ...option.RequestOption) (r *ChatService) {
 type ChatModel = string
 
 const (
+	ChatModelO1Preview           ChatModel = "o1-preview"
+	ChatModelO1Preview2024_09_12 ChatModel = "o1-preview-2024-09-12"
+	ChatModelO1Mini              ChatModel = "o1-mini"
+	ChatModelO1Mini2024_09_12    ChatModel = "o1-mini-2024-09-12"
 	ChatModelGPT4o               ChatModel = "gpt-4o"
-	ChatModelGPT4o2024_05_13     ChatModel = "gpt-4o-2024-05-13"
 	ChatModelGPT4o2024_08_06     ChatModel = "gpt-4o-2024-08-06"
+	ChatModelGPT4o2024_05_13     ChatModel = "gpt-4o-2024-05-13"
 	ChatModelChatgpt4oLatest     ChatModel = "chatgpt-4o-latest"
 	ChatModelGPT4oMini           ChatModel = "gpt-4o-mini"
 	ChatModelGPT4oMini2024_07_18 ChatModel = "gpt-4o-mini-2024-07-18"

chatcompletion.go

@@ -1483,13 +1483,17 @@ type ChatCompletionNewParams struct {
 	// returns the log probabilities of each output token returned in the `content` of
 	// `message`.
 	Logprobs param.Field[bool] `json:"logprobs"`
+	// An upper bound for the number of tokens that can be generated for a completion,
+	// including visible output tokens and
+	// [reasoning tokens](https://platform.openai.com/docs/guides/reasoning).
+	MaxCompletionTokens param.Field[int64] `json:"max_completion_tokens"`
 	// The maximum number of [tokens](/tokenizer) that can be generated in the chat
-	// completion.
+	// completion. This value can be used to control
+	// [costs](https://openai.com/api/pricing/) for text generated via API.
 	//
-	// The total length of input tokens and generated tokens is limited by the model's
-	// context length.
-	// [Example Python code](https://cookbook.openai.com/examples/how_to_count_tokens_with_tiktoken)
-	// for counting tokens.
+	// This value is now deprecated in favor of `max_completion_tokens`, and is not
+	// compatible with
+	// [o1 series models](https://platform.openai.com/docs/guides/reasoning).
 	MaxTokens param.Field[int64] `json:"max_tokens"`
 	// How many chat completion choices to generate for each input message. Note that
 	// you will be charged based on the number of generated tokens across all of the
@@ -1512,11 +1516,11 @@ type ChatCompletionNewParams struct {
 	// all GPT-3.5 Turbo models newer than `gpt-3.5-turbo-1106`.
 	//
 	// Setting to `{ "type": "json_schema", "json_schema": {...} }` enables Structured
-	// Outputs which guarantees the model will match your supplied JSON schema. Learn
-	// more in the
+	// Outputs which ensures the model will match your supplied JSON schema. Learn more
+	// in the
 	// [Structured Outputs guide](https://platform.openai.com/docs/guides/structured-outputs).
 	//
-	// Setting to `{ "type": "json_object" }` enables JSON mode, which guarantees the
+	// Setting to `{ "type": "json_object" }` enables JSON mode, which ensures the
 	// message the model generates is valid JSON.
 	//
 	// **Important:** when using JSON mode, you **must** also instruct the model to
@@ -1536,8 +1540,11 @@ type ChatCompletionNewParams struct {
 	// Specifies the latency tier to use for processing the request. This parameter is
 	// relevant for customers subscribed to the scale tier service:
 	//
-	//   - If set to 'auto', the system will utilize scale tier credits until they are
-	//     exhausted.
+	//   - If set to 'auto', and the Project is Scale tier enabled, the system will
+	//     utilize scale tier credits until they are exhausted.
+	//   - If set to 'auto', and the Project is not Scale tier enabled, the request will
+	//     be processed using the default service tier with a lower uptime SLA and no
+	//     latency guarentee.
 	//   - If set to 'default', the request will be processed using the default service
 	//     tier with a lower uptime SLA and no latency guarentee.
 	//   - When not set, the default behavior is 'auto'.
@@ -1655,11 +1662,11 @@ func (r ChatCompletionNewParamsFunction) MarshalJSON() (data []byte, err error)
 // all GPT-3.5 Turbo models newer than `gpt-3.5-turbo-1106`.
 //
 // Setting to `{ "type": "json_schema", "json_schema": {...} }` enables Structured
-// Outputs which guarantees the model will match your supplied JSON schema. Learn
-// more in the
+// Outputs which ensures the model will match your supplied JSON schema. Learn more
+// in the
 // [Structured Outputs guide](https://platform.openai.com/docs/guides/structured-outputs).
 //
-// Setting to `{ "type": "json_object" }` enables JSON mode, which guarantees the
+// Setting to `{ "type": "json_object" }` enables JSON mode, which ensures the
 // message the model generates is valid JSON.
 //
 // **Important:** when using JSON mode, you **must** also instruct the model to
@@ -1689,11 +1696,11 @@ func (r ChatCompletionNewParamsResponseFormat) ImplementsChatCompletionNewParams
 // all GPT-3.5 Turbo models newer than `gpt-3.5-turbo-1106`.
 //
 // Setting to `{ "type": "json_schema", "json_schema": {...} }` enables Structured
-// Outputs which guarantees the model will match your supplied JSON schema. Learn
-// more in the
+// Outputs which ensures the model will match your supplied JSON schema. Learn more
+// in the
 // [Structured Outputs guide](https://platform.openai.com/docs/guides/structured-outputs).
 //
-// Setting to `{ "type": "json_object" }` enables JSON mode, which guarantees the
+// Setting to `{ "type": "json_object" }` enables JSON mode, which ensures the
 // message the model generates is valid JSON.
 //
 // **Important:** when using JSON mode, you **must** also instruct the model to
@@ -1731,8 +1738,11 @@ func (r ChatCompletionNewParamsResponseFormatType) IsKnown() bool {
 // Specifies the latency tier to use for processing the request. This parameter is
 // relevant for customers subscribed to the scale tier service:
 //
-//   - If set to 'auto', the system will utilize scale tier credits until they are
-//     exhausted.
+//   - If set to 'auto', and the Project is Scale tier enabled, the system will
+//     utilize scale tier credits until they are exhausted.
+//   - If set to 'auto', and the Project is not Scale tier enabled, the request will
+//     be processed using the default service tier with a lower uptime SLA and no
+//     latency guarentee.
 //   - If set to 'default', the request will be processed using the default service
 //     tier with a lower uptime SLA and no latency guarentee.
 //   - When not set, the default behavior is 'auto'.

chatcompletion_test.go

@@ -32,7 +32,7 @@ func TestChatCompletionNewWithOptionalParams(t *testing.T) {
 			Role:    openai.F(openai.ChatCompletionSystemMessageParamRoleSystem),
 			Name:    openai.F("name"),
 		}}),
-		Model:            openai.F(openai.ChatModelGPT4o),
+		Model:            openai.F(openai.ChatModelO1Preview),
 		FrequencyPenalty: openai.F(-2.000000),
 		FunctionCall:     openai.F[openai.ChatCompletionNewParamsFunctionCallUnion](openai.ChatCompletionNewParamsFunctionCallString(openai.ChatCompletionNewParamsFunctionCallStringNone)),
 		Functions: openai.F([]openai.ChatCompletionNewParamsFunction{{
@@ -45,11 +45,12 @@ func TestChatCompletionNewWithOptionalParams(t *testing.T) {
 		LogitBias: openai.F(map[string]int64{
 			"foo": int64(0),
 		}),
-		Logprobs:          openai.F(true),
-		MaxTokens:         openai.F(int64(0)),
-		N:                 openai.F(int64(1)),
-		ParallelToolCalls: openai.F(true),
-		PresencePenalty:   openai.F(-2.000000),
+		Logprobs:            openai.F(true),
+		MaxCompletionTokens: openai.F(int64(0)),
+		MaxTokens:           openai.F(int64(0)),
+		N:                   openai.F(int64(1)),
+		ParallelToolCalls:   openai.F(true),
+		PresencePenalty:     openai.F(-2.000000),
 		ResponseFormat: openai.F[openai.ChatCompletionNewParamsResponseFormatUnion](shared.ResponseFormatTextParam{
 			Type: openai.F(shared.ResponseFormatTextTypeText),
 		}),

client_test.go

@@ -41,7 +41,7 @@ func TestUserAgentHeader(t *testing.T) {
 			Role:    openai.F(openai.ChatCompletionUserMessageParamRoleUser),
 			Content: openai.F([]openai.ChatCompletionContentPartUnionParam{openai.ChatCompletionContentPartTextParam{Text: openai.F("text"), Type: openai.F(openai.ChatCompletionContentPartTextTypeText)}}),
 		}}),
-		Model: openai.F(openai.ChatModelGPT4o),
+		Model: openai.F(openai.ChatModelO1Preview),
 	})
 	if userAgent != fmt.Sprintf("OpenAI/Go %s", internal.PackageVersion) {
 		t.Errorf("Expected User-Agent to be correct, but got: %#v", userAgent)
@@ -70,7 +70,7 @@ func TestRetryAfter(t *testing.T) {
 			Role:    openai.F(openai.ChatCompletionUserMessageParamRoleUser),
 			Content: openai.F([]openai.ChatCompletionContentPartUnionParam{openai.ChatCompletionContentPartTextParam{Text: openai.F("text"), Type: openai.F(openai.ChatCompletionContentPartTextTypeText)}}),
 		}}),
-		Model: openai.F(openai.ChatModelGPT4o),
+		Model: openai.F(openai.ChatModelO1Preview),
 	})
 	if err == nil || res != nil {
 		t.Error("Expected there to be a cancel error and for the response to be nil")
@@ -102,7 +102,7 @@ func TestRetryAfterMs(t *testing.T) {
 			Role:    openai.F(openai.ChatCompletionUserMessageParamRoleUser),
 			Content: openai.F([]openai.ChatCompletionContentPartUnionParam{openai.ChatCompletionContentPartTextParam{Text: openai.F("text"), Type: openai.F(openai.ChatCompletionContentPartTextTypeText)}}),
 		}}),
-		Model: openai.F(openai.ChatModelGPT4o),
+		Model: openai.F(openai.ChatModelO1Preview),
 	})
 	if err == nil || res != nil {
 		t.Error("Expected there to be a cancel error and for the response to be nil")
@@ -130,7 +130,7 @@ func TestContextCancel(t *testing.T) {
 			Role:    openai.F(openai.ChatCompletionUserMessageParamRoleUser),
 			Content: openai.F([]openai.ChatCompletionContentPartUnionParam{openai.ChatCompletionContentPartTextParam{Text: openai.F("text"), Type: openai.F(openai.ChatCompletionContentPartTextTypeText)}}),
 		}}),
-		Model: openai.F(openai.ChatModelGPT4o),
+		Model: openai.F(openai.ChatModelO1Preview),
 	})
 	if err == nil || res != nil {
 		t.Error("Expected there to be a cancel error and for the response to be nil")
@@ -155,7 +155,7 @@ func TestContextCancelDelay(t *testing.T) {
 			Role:    openai.F(openai.ChatCompletionUserMessageParamRoleUser),
 			Content: openai.F([]openai.ChatCompletionContentPartUnionParam{openai.ChatCompletionContentPartTextParam{Text: openai.F("text"), Type: openai.F(openai.ChatCompletionContentPartTextTypeText)}}),
 		}}),
-		Model: openai.F(openai.ChatModelGPT4o),
+		Model: openai.F(openai.ChatModelO1Preview),
 	})
 	if err == nil || res != nil {
 		t.Error("expected there to be a cancel error and for the response to be nil")
@@ -186,7 +186,7 @@ func TestContextDeadline(t *testing.T) {
 				Role:    openai.F(openai.ChatCompletionUserMessageParamRoleUser),
 				Content: openai.F([]openai.ChatCompletionContentPartUnionParam{openai.ChatCompletionContentPartTextParam{Text: openai.F("text"), Type: openai.F(openai.ChatCompletionContentPartTextTypeText)}}),
 			}}),
-			Model: openai.F(openai.ChatModelGPT4o),
+			Model: openai.F(openai.ChatModelO1Preview),
 		})
 		if err == nil || res != nil {
 			t.Error("expected there to be a deadline error and for the response to be nil")

completion.go

@@ -197,17 +197,20 @@ type CompletionUsage struct {
 	// Number of tokens in the prompt.
 	PromptTokens int64 `json:"prompt_tokens,required"`
 	// Total number of tokens used in the request (prompt + completion).
-	TotalTokens int64               `json:"total_tokens,required"`
-	JSON        completionUsageJSON `json:"-"`
+	TotalTokens int64 `json:"total_tokens,required"`
+	// Breakdown of tokens used in a completion.
+	CompletionTokensDetails CompletionUsageCompletionTokensDetails `json:"completion_tokens_details"`
+	JSON                    completionUsageJSON                    `json:"-"`
 }
 
 // completionUsageJSON contains the JSON metadata for the struct [CompletionUsage]
 type completionUsageJSON struct {
-	CompletionTokens apijson.Field
-	PromptTokens     apijson.Field
-	TotalTokens      apijson.Field
-	raw              string
-	ExtraFields      map[string]apijson.Field
+	CompletionTokens        apijson.Field
+	PromptTokens            apijson.Field
+	TotalTokens             apijson.Field
+	CompletionTokensDetails apijson.Field
+	raw                     string
+	ExtraFields             map[string]apijson.Field
 }
 
 func (r *CompletionUsage) UnmarshalJSON(data []byte) (err error) {
@@ -218,6 +221,29 @@ func (r completionUsageJSON) RawJSON() string {
 	return r.raw
 }
 
+// Breakdown of tokens used in a completion.
+type CompletionUsageCompletionTokensDetails struct {
+	// Tokens generated by the model for reasoning.
+	ReasoningTokens int64                                      `json:"reasoning_tokens"`
+	JSON            completionUsageCompletionTokensDetailsJSON `json:"-"`
+}
+
+// completionUsageCompletionTokensDetailsJSON contains the JSON metadata for the
+// struct [CompletionUsageCompletionTokensDetails]
+type completionUsageCompletionTokensDetailsJSON struct {
+	ReasoningTokens apijson.Field
+	raw             string
+	ExtraFields     map[string]apijson.Field
+}
+
+func (r *CompletionUsageCompletionTokensDetails) UnmarshalJSON(data []byte) (err error) {
+	return apijson.UnmarshalRoot(data, r)
+}
+
+func (r completionUsageCompletionTokensDetailsJSON) RawJSON() string {
+	return r.raw
+}
+
 type CompletionNewParams struct {
 	// ID of the model to use. You can use the
 	// [List models](https://platform.openai.com/docs/api-reference/models/list) API to

finetuningjob.go

@@ -524,7 +524,7 @@ type FineTuningJobNewParams struct {
 	// job parameters should produce the same results, but may differ in rare cases. If
 	// a seed is not specified, one will be generated for you.
 	Seed param.Field[int64] `json:"seed"`
-	// A string of up to 18 characters that will be added to your fine-tuned model
+	// A string of up to 64 characters that will be added to your fine-tuned model
 	// name.
 	//
 	// For example, a `suffix` of "custom-model-name" would produce a model name like

usage_test.go

@@ -29,7 +29,7 @@ func TestUsage(t *testing.T) {
 			Role:    openai.F(openai.ChatCompletionUserMessageParamRoleUser),
 			Content: openai.F([]openai.ChatCompletionContentPartUnionParam{openai.ChatCompletionContentPartTextParam{Text: openai.F("text"), Type: openai.F(openai.ChatCompletionContentPartTextTypeText)}}),
 		}}),
-		Model: openai.F(openai.ChatModelGPT4o),
+		Model: openai.F(openai.ChatModelO1Preview),
 	})
 	if err != nil {
 		t.Error(err)