feat: able to use ENFORCE_INCLUDE_USAGE to enforce include usage in response

* fix:修复在渠道配置中设置模型重定向时，temperature为0被忽略的问题

* fix: set optional fields to pointer type

---------

Co-authored-by: JustSong <songquanpeng@foxmail.com>

README.md

@@ -90,6 +90,7 @@ _✨ 通过标准的 OpenAI API 格式访问所有的大模型，开箱即用
    + [x] [together.ai](https://www.together.ai/)
    + [x] [novita.ai](https://www.novita.ai/)
    + [x] [硅基流动 SiliconCloud](https://siliconflow.cn/siliconcloud)
+   + [x] [xAI](https://x.ai/)
 2. 支持配置镜像以及众多[第三方代理服务](https://iamazing.cn/page/openai-api-third-party-services)。
 3. 支持通过**负载均衡**的方式访问多个渠道。
 4. 支持 **stream 模式**，可以通过流式传输实现打字机效果。
@@ -399,6 +400,7 @@ graph LR
 26. `METRIC_SUCCESS_RATE_THRESHOLD`：请求成功率阈值，默认为 `0.8`。
 27. `INITIAL_ROOT_TOKEN`：如果设置了该值，则在系统首次启动时会自动创建一个值为该环境变量值的 root 用户令牌。
 28. `INITIAL_ROOT_ACCESS_TOKEN`：如果设置了该值，则在系统首次启动时会自动创建一个值为该环境变量的 root 用户创建系统管理令牌。
+29. `ENFORCE_INCLUDE_USAGE`：是否强制在 stream 模型下返回 usage，默认不开启，可选值为 `true` 和 `false`。
 
 ### 命令行参数
 1. `--port <port_number>`: 指定服务器监听的端口号，默认为 `3000`。

common/config/config.go

@@ -160,3 +160,5 @@ var OnlyOneLogFile = env.Bool("ONLY_ONE_LOG_FILE", false)
 var RelayProxy = env.String("RELAY_PROXY", "")
 var UserContentRequestProxy = env.String("USER_CONTENT_REQUEST_PROXY", "")
 var UserContentRequestTimeout = env.Int("USER_CONTENT_REQUEST_TIMEOUT", 30)
+
+var EnforceIncludeUsage = env.Bool("ENFORCE_INCLUDE_USAGE", false)

common/helper/helper.go

@@ -137,3 +137,23 @@ func String2Int(str string) int {
 	}
 	return num
 }
+
+func Float64PtrMax(p *float64, maxValue float64) *float64 {
+	if p == nil {
+		return nil
+	}
+	if *p > maxValue {
+		return &maxValue
+	}
+	return p
+}
+
+func Float64PtrMin(p *float64, minValue float64) *float64 {
+	if p == nil {
+		return nil
+	}
+	if *p < minValue {
+		return &minValue
+	}
+	return p
+}

controller/channel-test.go

@@ -76,9 +76,9 @@ func testChannel(channel *model.Channel, request *relaymodel.GeneralOpenAIReques
 		if len(modelNames) > 0 {
 			modelName = modelNames[0]
 		}
-		if modelMap != nil && modelMap[modelName] != "" {
-			modelName = modelMap[modelName]
-		}
+	}
+	if modelMap != nil && modelMap[modelName] != "" {
+		modelName = modelMap[modelName]
 	}
 	meta.OriginModelName, meta.ActualModelName = request.Model, modelName
 	request.Model = modelName

relay/adaptor/ali/main.go

@@ -36,9 +36,7 @@ func ConvertRequest(request model.GeneralOpenAIRequest) *ChatRequest {
 		enableSearch = true
 		aliModel = strings.TrimSuffix(aliModel, EnableSearchModelSuffix)
 	}
-	if request.TopP >= 1 {
-		request.TopP = 0.9999
-	}
+	request.TopP = helper.Float64PtrMax(request.TopP, 0.9999)
 	return &ChatRequest{
 		Model: aliModel,
 		Input: Input{

relay/adaptor/ali/model.go

@@ -16,13 +16,13 @@ type Input struct {
 }
 
 type Parameters struct {
-	TopP              float64      `json:"top_p,omitempty"`
+	TopP              *float64     `json:"top_p,omitempty"`
 	TopK              int          `json:"top_k,omitempty"`
 	Seed              uint64       `json:"seed,omitempty"`
 	EnableSearch      bool         `json:"enable_search,omitempty"`
 	IncrementalOutput bool         `json:"incremental_output,omitempty"`
 	MaxTokens         int          `json:"max_tokens,omitempty"`
-	Temperature       float64      `json:"temperature,omitempty"`
+	Temperature       *float64     `json:"temperature,omitempty"`
 	ResultFormat      string       `json:"result_format,omitempty"`
 	Tools             []model.Tool `json:"tools,omitempty"`
 }

relay/adaptor/anthropic/constants.go

@@ -3,7 +3,11 @@ package anthropic
 var ModelList = []string{
 	"claude-instant-1.2", "claude-2.0", "claude-2.1",
 	"claude-3-haiku-20240307",
+	"claude-3-5-haiku-20241022",
 	"claude-3-sonnet-20240229",
 	"claude-3-opus-20240229",
 	"claude-3-5-sonnet-20240620",
+	"claude-3-5-sonnet-20241022",
+	"claude-3-5-sonnet-latest",
+	"claude-3-5-haiku-20241022",
 }

relay/adaptor/anthropic/model.go

@@ -48,8 +48,8 @@ type Request struct {
 	MaxTokens     int       `json:"max_tokens,omitempty"`
 	StopSequences []string  `json:"stop_sequences,omitempty"`
 	Stream        bool      `json:"stream,omitempty"`
-	Temperature   float64   `json:"temperature,omitempty"`
-	TopP          float64   `json:"top_p,omitempty"`
+	Temperature   *float64  `json:"temperature,omitempty"`
+	TopP          *float64  `json:"top_p,omitempty"`
 	TopK          int       `json:"top_k,omitempty"`
 	Tools         []Tool    `json:"tools,omitempty"`
 	ToolChoice    any       `json:"tool_choice,omitempty"`

relay/adaptor/aws/claude/main.go

@@ -29,10 +29,13 @@ var AwsModelIDMap = map[string]string{
 	"claude-instant-1.2":         "anthropic.claude-instant-v1",
 	"claude-2.0":                 "anthropic.claude-v2",
 	"claude-2.1":                 "anthropic.claude-v2:1",
-	"claude-3-sonnet-20240229":   "anthropic.claude-3-sonnet-20240229-v1:0",
-	"claude-3-5-sonnet-20240620": "anthropic.claude-3-5-sonnet-20240620-v1:0",
-	"claude-3-opus-20240229":     "anthropic.claude-3-opus-20240229-v1:0",
 	"claude-3-haiku-20240307":    "anthropic.claude-3-haiku-20240307-v1:0",
+	"claude-3-sonnet-20240229":   "anthropic.claude-3-sonnet-20240229-v1:0",
+	"claude-3-opus-20240229":     "anthropic.claude-3-opus-20240229-v1:0",
+	"claude-3-5-sonnet-20240620": "anthropic.claude-3-5-sonnet-20240620-v1:0",
+	"claude-3-5-sonnet-20241022": "anthropic.claude-3-5-sonnet-20241022-v2:0",
+	"claude-3-5-sonnet-latest":   "anthropic.claude-3-5-sonnet-20241022-v2:0",
+	"claude-3-5-haiku-20241022":  "anthropic.claude-3-5-haiku-20241022-v1:0",
 }
 
 func awsModelID(requestModel string) (string, error) {

relay/adaptor/aws/claude/model.go

@@ -11,8 +11,8 @@ type Request struct {
 	Messages         []anthropic.Message `json:"messages"`
 	System           string              `json:"system,omitempty"`
 	MaxTokens        int                 `json:"max_tokens,omitempty"`
-	Temperature      float64             `json:"temperature,omitempty"`
-	TopP             float64             `json:"top_p,omitempty"`
+	Temperature      *float64            `json:"temperature,omitempty"`
+	TopP             *float64            `json:"top_p,omitempty"`
 	TopK             int                 `json:"top_k,omitempty"`
 	StopSequences    []string            `json:"stop_sequences,omitempty"`
 	Tools            []anthropic.Tool    `json:"tools,omitempty"`

relay/adaptor/aws/llama3/model.go

@@ -4,10 +4,10 @@ package aws
 //
 // https://docs.aws.amazon.com/bedrock/latest/userguide/model-parameters-meta.html
 type Request struct {
-	Prompt      string  `json:"prompt"`
-	MaxGenLen   int     `json:"max_gen_len,omitempty"`
-	Temperature float64 `json:"temperature,omitempty"`
-	TopP        float64 `json:"top_p,omitempty"`
+	Prompt      string   `json:"prompt"`
+	MaxGenLen   int      `json:"max_gen_len,omitempty"`
+	Temperature *float64 `json:"temperature,omitempty"`
+	TopP        *float64 `json:"top_p,omitempty"`
 }
 
 // Response is the response from AWS Llama3

relay/adaptor/baidu/main.go

@@ -35,9 +35,9 @@ type Message struct {
 
 type ChatRequest struct {
 	Messages        []Message `json:"messages"`
-	Temperature     float64   `json:"temperature,omitempty"`
-	TopP            float64   `json:"top_p,omitempty"`
-	PenaltyScore    float64   `json:"penalty_score,omitempty"`
+	Temperature     *float64  `json:"temperature,omitempty"`
+	TopP            *float64  `json:"top_p,omitempty"`
+	PenaltyScore    *float64  `json:"penalty_score,omitempty"`
 	Stream          bool      `json:"stream,omitempty"`
 	System          string    `json:"system,omitempty"`
 	DisableSearch   bool      `json:"disable_search,omitempty"`

relay/adaptor/cloudflare/model.go

@@ -9,5 +9,5 @@ type Request struct {
 	Prompt      string          `json:"prompt,omitempty"`
 	Raw         bool            `json:"raw,omitempty"`
 	Stream      bool            `json:"stream,omitempty"`
-	Temperature float64         `json:"temperature,omitempty"`
+	Temperature *float64        `json:"temperature,omitempty"`
 }

relay/adaptor/cohere/main.go

@@ -43,7 +43,7 @@ func ConvertRequest(textRequest model.GeneralOpenAIRequest) *Request {
 		K:                textRequest.TopK,
 		Stream:           textRequest.Stream,
 		FrequencyPenalty: textRequest.FrequencyPenalty,
-		PresencePenalty:  textRequest.FrequencyPenalty,
+		PresencePenalty:  textRequest.PresencePenalty,
 		Seed:             int(textRequest.Seed),
 	}
 	if cohereRequest.Model == "" {

relay/adaptor/cohere/model.go

@@ -10,15 +10,15 @@ type Request struct {
 	PromptTruncation string        `json:"prompt_truncation,omitempty"` // 默认值为"AUTO"
 	Connectors       []Connector   `json:"connectors,omitempty"`
 	Documents        []Document    `json:"documents,omitempty"`
-	Temperature      float64       `json:"temperature,omitempty"` // 默认值为0.3
+	Temperature      *float64      `json:"temperature,omitempty"` // 默认值为0.3
 	MaxTokens        int           `json:"max_tokens,omitempty"`
 	MaxInputTokens   int           `json:"max_input_tokens,omitempty"`
 	K                int           `json:"k,omitempty"` // 默认值为0
-	P                float64       `json:"p,omitempty"` // 默认值为0.75
+	P                *float64      `json:"p,omitempty"` // 默认值为0.75
 	Seed             int           `json:"seed,omitempty"`
 	StopSequences    []string      `json:"stop_sequences,omitempty"`
-	FrequencyPenalty float64       `json:"frequency_penalty,omitempty"` // 默认值为0.0
-	PresencePenalty  float64       `json:"presence_penalty,omitempty"`  // 默认值为0.0
+	FrequencyPenalty *float64      `json:"frequency_penalty,omitempty"` // 默认值为0.0
+	PresencePenalty  *float64      `json:"presence_penalty,omitempty"`  // 默认值为0.0
 	Tools            []Tool        `json:"tools,omitempty"`
 	ToolResults      []ToolResult  `json:"tool_results,omitempty"`
 }

relay/adaptor/gemini/main.go

@@ -4,11 +4,12 @@ import (
 	"bufio"
 	"encoding/json"
 	"fmt"
-	"github.com/songquanpeng/one-api/common/render"
 	"io"
 	"net/http"
 	"strings"
 
+	"github.com/songquanpeng/one-api/common/render"
+
 	"github.com/songquanpeng/one-api/common"
 	"github.com/songquanpeng/one-api/common/config"
 	"github.com/songquanpeng/one-api/common/helper"
@@ -28,6 +29,11 @@ const (
 	VisionMaxImageNum = 16
 )
 
+var mimeTypeMap = map[string]string{
+	"json_object": "application/json",
+	"text":        "text/plain",
+}
+
 // Setting safety to the lowest possible values since Gemini is already powerless enough
 func ConvertRequest(textRequest model.GeneralOpenAIRequest) *ChatRequest {
 	geminiRequest := ChatRequest{
@@ -56,6 +62,15 @@ func ConvertRequest(textRequest model.GeneralOpenAIRequest) *ChatRequest {
 			MaxOutputTokens: textRequest.MaxTokens,
 		},
 	}
+	if textRequest.ResponseFormat != nil {
+		if mimeType, ok := mimeTypeMap[textRequest.ResponseFormat.Type]; ok {
+			geminiRequest.GenerationConfig.ResponseMimeType = mimeType
+		}
+		if textRequest.ResponseFormat.JsonSchema != nil {
+			geminiRequest.GenerationConfig.ResponseSchema = textRequest.ResponseFormat.JsonSchema.Schema
+			geminiRequest.GenerationConfig.ResponseMimeType = mimeTypeMap["json_object"]
+		}
+	}
 	if textRequest.Tools != nil {
 		functions := make([]model.Function, 0, len(textRequest.Tools))
 		for _, tool := range textRequest.Tools {

relay/adaptor/gemini/model.go

@@ -65,10 +65,12 @@ type ChatTools struct {
 }
 
 type ChatGenerationConfig struct {
-	Temperature     float64  `json:"temperature,omitempty"`
-	TopP            float64  `json:"topP,omitempty"`
-	TopK            float64  `json:"topK,omitempty"`
-	MaxOutputTokens int      `json:"maxOutputTokens,omitempty"`
-	CandidateCount  int      `json:"candidateCount,omitempty"`
-	StopSequences   []string `json:"stopSequences,omitempty"`
+	ResponseMimeType string   `json:"responseMimeType,omitempty"`
+	ResponseSchema   any      `json:"responseSchema,omitempty"`
+	Temperature      *float64 `json:"temperature,omitempty"`
+	TopP             *float64 `json:"topP,omitempty"`
+	TopK             float64  `json:"topK,omitempty"`
+	MaxOutputTokens  int      `json:"maxOutputTokens,omitempty"`
+	CandidateCount   int      `json:"candidateCount,omitempty"`
+	StopSequences    []string `json:"stopSequences,omitempty"`
 }

relay/adaptor/groq/constants.go

@@ -4,14 +4,24 @@ package groq
 
 var ModelList = []string{
 	"gemma-7b-it",
-	"mixtral-8x7b-32768",
-	"llama3-8b-8192",
-	"llama3-70b-8192",
 	"gemma2-9b-it",
-	"llama-3.1-405b-reasoning",
 	"llama-3.1-70b-versatile",
 	"llama-3.1-8b-instant",
+	"llama-3.2-11b-text-preview",
+	"llama-3.2-11b-vision-preview",
+	"llama-3.2-1b-preview",
+	"llama-3.2-3b-preview",
+	"llama-3.2-11b-vision-preview",
+	"llama-3.2-90b-text-preview",
+	"llama-3.2-90b-vision-preview",
+	"llama-guard-3-8b",
+	"llama3-70b-8192",
+	"llama3-8b-8192",
 	"llama3-groq-70b-8192-tool-use-preview",
 	"llama3-groq-8b-8192-tool-use-preview",
+	"llava-v1.5-7b-4096-preview",
+	"mixtral-8x7b-32768",
+	"distil-whisper-large-v3-en",
 	"whisper-large-v3",
+	"whisper-large-v3-turbo",
 }

relay/adaptor/ollama/model.go

@@ -1,14 +1,14 @@
 package ollama
 
 type Options struct {
-	Seed             int     `json:"seed,omitempty"`
-	Temperature      float64 `json:"temperature,omitempty"`
-	TopK             int     `json:"top_k,omitempty"`
-	TopP             float64 `json:"top_p,omitempty"`
-	FrequencyPenalty float64 `json:"frequency_penalty,omitempty"`
-	PresencePenalty  float64 `json:"presence_penalty,omitempty"`
-	NumPredict  	 int 	 `json:"num_predict,omitempty"`
-	NumCtx  	 int 	 `json:"num_ctx,omitempty"`
+	Seed             int      `json:"seed,omitempty"`
+	Temperature      *float64 `json:"temperature,omitempty"`
+	TopK             int      `json:"top_k,omitempty"`
+	TopP             *float64 `json:"top_p,omitempty"`
+	FrequencyPenalty *float64 `json:"frequency_penalty,omitempty"`
+	PresencePenalty  *float64 `json:"presence_penalty,omitempty"`
+	NumPredict       int      `json:"num_predict,omitempty"`
+	NumCtx           int      `json:"num_ctx,omitempty"`
 }
 
 type Message struct {

relay/adaptor/openai/adaptor.go

@@ -75,6 +75,13 @@ func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.G
 	if request == nil {
 		return nil, errors.New("request is nil")
 	}
+	if request.Stream {
+		// always return usage in stream mode
+		if request.StreamOptions == nil {
+			request.StreamOptions = &model.StreamOptions{}
+		}
+		request.StreamOptions.IncludeUsage = true
+	}
 	return request, nil
 }
 

relay/adaptor/openai/compatible.go

@@ -11,9 +11,10 @@ import (
 	"github.com/songquanpeng/one-api/relay/adaptor/mistral"
 	"github.com/songquanpeng/one-api/relay/adaptor/moonshot"
 	"github.com/songquanpeng/one-api/relay/adaptor/novita"
+	"github.com/songquanpeng/one-api/relay/adaptor/siliconflow"
 	"github.com/songquanpeng/one-api/relay/adaptor/stepfun"
 	"github.com/songquanpeng/one-api/relay/adaptor/togetherai"
-	"github.com/songquanpeng/one-api/relay/adaptor/siliconflow"
+	"github.com/songquanpeng/one-api/relay/adaptor/xai"
 	"github.com/songquanpeng/one-api/relay/channeltype"
 )
 
@@ -32,6 +33,7 @@ var CompatibleChannels = []int{
 	channeltype.TogetherAI,
 	channeltype.Novita,
 	channeltype.SiliconFlow,
+	channeltype.XAI,
 }
 
 func GetCompatibleChannelMeta(channelType int) (string, []string) {
@@ -64,6 +66,8 @@ func GetCompatibleChannelMeta(channelType int) (string, []string) {
 		return "novita", novita.ModelList
 	case channeltype.SiliconFlow:
 		return "siliconflow", siliconflow.ModelList
+	case channeltype.XAI:
+		return "xai", xai.ModelList
 	default:
 		return "openai", ModelList
 	}

relay/adaptor/palm/model.go

@@ -19,11 +19,11 @@ type Prompt struct {
 }
 
 type ChatRequest struct {
-	Prompt         Prompt  `json:"prompt"`
-	Temperature    float64 `json:"temperature,omitempty"`
-	CandidateCount int     `json:"candidateCount,omitempty"`
-	TopP           float64 `json:"topP,omitempty"`
-	TopK           int     `json:"topK,omitempty"`
+	Prompt         Prompt   `json:"prompt"`
+	Temperature    *float64 `json:"temperature,omitempty"`
+	CandidateCount int      `json:"candidateCount,omitempty"`
+	TopP           *float64 `json:"topP,omitempty"`
+	TopK           int      `json:"topK,omitempty"`
 }
 
 type Error struct {

relay/adaptor/tencent/main.go

@@ -39,8 +39,8 @@ func ConvertRequest(request model.GeneralOpenAIRequest) *ChatRequest {
 		Model:       &request.Model,
 		Stream:      &request.Stream,
 		Messages:    messages,
-		TopP:        &request.TopP,
-		Temperature: &request.Temperature,
+		TopP:        request.TopP,
+		Temperature: request.Temperature,
 	}
 }
 

relay/adaptor/vertexai/claude/adapter.go

@@ -13,7 +13,12 @@ import (
 )
 
 var ModelList = []string{
-	"claude-3-haiku@20240307", "claude-3-opus@20240229", "claude-3-5-sonnet@20240620", "claude-3-sonnet@20240229",
+	"claude-3-haiku@20240307",
+	"claude-3-sonnet@20240229",
+	"claude-3-opus@20240229",
+	"claude-3-5-sonnet@20240620",
+	"claude-3-5-sonnet-v2@20241022",
+	"claude-3-5-haiku@20241022",
 }
 
 const anthropicVersion = "vertex-2023-10-16"

relay/adaptor/vertexai/claude/model.go

@@ -11,8 +11,8 @@ type Request struct {
 	MaxTokens     int                 `json:"max_tokens,omitempty"`
 	StopSequences []string            `json:"stop_sequences,omitempty"`
 	Stream        bool                `json:"stream,omitempty"`
-	Temperature   float64             `json:"temperature,omitempty"`
-	TopP          float64             `json:"top_p,omitempty"`
+	Temperature   *float64            `json:"temperature,omitempty"`
+	TopP          *float64            `json:"top_p,omitempty"`
 	TopK          int                 `json:"top_k,omitempty"`
 	Tools         []anthropic.Tool    `json:"tools,omitempty"`
 	ToolChoice    any                 `json:"tool_choice,omitempty"`

relay/adaptor/vertexai/gemini/adapter.go

@@ -15,7 +15,7 @@ import (
 )
 
 var ModelList = []string{
-	"gemini-1.5-pro-001", "gemini-1.5-flash-001", "gemini-pro", "gemini-pro-vision",
+	"gemini-1.5-pro-001", "gemini-1.5-flash-001", "gemini-pro", "gemini-pro-vision", "gemini-1.5-pro-002", "gemini-1.5-flash-002", 
 }
 
 type Adaptor struct {

relay/adaptor/xai/constants.go

@@ -0,0 +1,5 @@
+package xai
+
+var ModelList = []string{
+	"grok-beta",
+}

relay/adaptor/xunfei/constants.go

@@ -7,5 +7,6 @@ var ModelList = []string{
 	"SparkDesk-v3.1",
 	"SparkDesk-v3.1-128K",
 	"SparkDesk-v3.5",
+	"SparkDesk-v3.5-32K",
 	"SparkDesk-v4.0",
 }

relay/adaptor/xunfei/main.go

@@ -283,7 +283,7 @@ func parseAPIVersionByModelName(modelName string) string {
 func apiVersion2domain(apiVersion string) string {
 	switch apiVersion {
 	case "v1.1":
-		return "general"
+		return "lite"
 	case "v2.1":
 		return "generalv2"
 	case "v3.1":
@@ -292,6 +292,8 @@ func apiVersion2domain(apiVersion string) string {
 		return "pro-128k"
 	case "v3.5":
 		return "generalv3.5"
+	case "v3.5-32K":
+		return "max-32k"
 	case "v4.0":
 		return "4.0Ultra"
 	}
@@ -303,7 +305,10 @@ func getXunfeiAuthUrl(apiVersion string, apiKey string, apiSecret string) (strin
 	domain := apiVersion2domain(apiVersion)
 	switch apiVersion {
 	case "v3.1-128K":
-		authUrl = buildXunfeiAuthUrl(fmt.Sprintf("wss://spark-api.xf-yun.com/%s/pro-128k", apiVersion), apiKey, apiSecret)
+		authUrl = buildXunfeiAuthUrl(fmt.Sprintf("wss://spark-api.xf-yun.com/chat/pro-128k"), apiKey, apiSecret)
+		break
+	case "v3.5-32K":
+		authUrl = buildXunfeiAuthUrl(fmt.Sprintf("wss://spark-api.xf-yun.com/chat/max-32k"), apiKey, apiSecret)
 		break
 	default:
 		authUrl = buildXunfeiAuthUrl(fmt.Sprintf("wss://spark-api.xf-yun.com/%s/chat", apiVersion), apiKey, apiSecret)

relay/adaptor/xunfei/model.go

@@ -19,11 +19,11 @@ type ChatRequest struct {
 	} `json:"header"`
 	Parameter struct {
 		Chat struct {
-			Domain      string  `json:"domain,omitempty"`
-			Temperature float64 `json:"temperature,omitempty"`
-			TopK        int     `json:"top_k,omitempty"`
-			MaxTokens   int     `json:"max_tokens,omitempty"`
-			Auditing    bool    `json:"auditing,omitempty"`
+			Domain      string   `json:"domain,omitempty"`
+			Temperature *float64 `json:"temperature,omitempty"`
+			TopK        int      `json:"top_k,omitempty"`
+			MaxTokens   int      `json:"max_tokens,omitempty"`
+			Auditing    bool     `json:"auditing,omitempty"`
 		} `json:"chat"`
 	} `json:"parameter"`
 	Payload struct {

relay/adaptor/zhipu/adaptor.go

@@ -4,13 +4,13 @@ import (
 	"errors"
 	"fmt"
 	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common/helper"
 	"github.com/songquanpeng/one-api/relay/adaptor"
 	"github.com/songquanpeng/one-api/relay/adaptor/openai"
 	"github.com/songquanpeng/one-api/relay/meta"
 	"github.com/songquanpeng/one-api/relay/model"
 	"github.com/songquanpeng/one-api/relay/relaymode"
 	"io"
-	"math"
 	"net/http"
 	"strings"
 )
@@ -65,13 +65,13 @@ func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.G
 		baiduEmbeddingRequest, err := ConvertEmbeddingRequest(*request)
 		return baiduEmbeddingRequest, err
 	default:
-		// TopP (0.0, 1.0)
-		request.TopP = math.Min(0.99, request.TopP)
-		request.TopP = math.Max(0.01, request.TopP)
+		// TopP [0.0, 1.0]
+		request.TopP = helper.Float64PtrMax(request.TopP, 1)
+		request.TopP = helper.Float64PtrMin(request.TopP, 0)
 
-		// Temperature (0.0, 1.0)
-		request.Temperature = math.Min(0.99, request.Temperature)
-		request.Temperature = math.Max(0.01, request.Temperature)
+		// Temperature [0.0, 1.0]
+		request.Temperature = helper.Float64PtrMax(request.Temperature, 1)
+		request.Temperature = helper.Float64PtrMin(request.Temperature, 0)
 		a.SetVersionByModeName(request.Model)
 		if a.APIVersion == "v4" {
 			return request, nil

relay/adaptor/zhipu/model.go

@@ -12,8 +12,8 @@ type Message struct {
 
 type Request struct {
 	Prompt      []Message `json:"prompt"`
-	Temperature float64   `json:"temperature,omitempty"`
-	TopP        float64   `json:"top_p,omitempty"`
+	Temperature *float64  `json:"temperature,omitempty"`
+	TopP        *float64  `json:"top_p,omitempty"`
 	RequestId   string    `json:"request_id,omitempty"`
 	Incremental bool      `json:"incremental,omitempty"`
 }

relay/billing/ratio/model.go

@@ -79,8 +79,10 @@ var ModelRatio = map[string]float64{
 	"claude-2.0":                 8.0 / 1000 * USD,
 	"claude-2.1":                 8.0 / 1000 * USD,
 	"claude-3-haiku-20240307":    0.25 / 1000 * USD,
+	"claude-3-5-haiku-20241022":  1.0 / 1000 * USD,
 	"claude-3-sonnet-20240229":   3.0 / 1000 * USD,
 	"claude-3-5-sonnet-20240620": 3.0 / 1000 * USD,
+	"claude-3-5-sonnet-20241022": 3.0 / 1000 * USD,
 	"claude-3-opus-20240229":     15.0 / 1000 * USD,
 	// https://cloud.baidu.com/doc/WENXINWORKSHOP/s/hlrk4akp7
 	"ERNIE-4.0-8K":       0.120 * RMB,
@@ -130,6 +132,7 @@ var ModelRatio = map[string]float64{
 	"SparkDesk-v3.1":            1.2858, // ￥0.018 / 1k tokens
 	"SparkDesk-v3.1-128K":       1.2858, // ￥0.018 / 1k tokens
 	"SparkDesk-v3.5":            1.2858, // ￥0.018 / 1k tokens
+	"SparkDesk-v3.5-32K":        1.2858, // ￥0.018 / 1k tokens
 	"SparkDesk-v4.0":            1.2858, // ￥0.018 / 1k tokens
 	"360GPT_S2_V9":              0.8572, // ¥0.012 / 1k tokens
 	"embedding-bert-512-v1":     0.0715, // ¥0.001 / 1k tokens
@@ -161,15 +164,21 @@ var ModelRatio = map[string]float64{
 	"mistral-embed":         0.1 / 1000 * USD,
 	// https://wow.groq.com/#:~:text=inquiries%C2%A0here.-,Model,-Current%20Speed
 	"gemma-7b-it":                           0.07 / 1000000 * USD,
-	"mixtral-8x7b-32768":                    0.24 / 1000000 * USD,
-	"llama3-8b-8192":                        0.05 / 1000000 * USD,
-	"llama3-70b-8192":                       0.59 / 1000000 * USD,
 	"gemma2-9b-it":                          0.20 / 1000000 * USD,
-	"llama-3.1-405b-reasoning":              0.89 / 1000000 * USD,
 	"llama-3.1-70b-versatile":               0.59 / 1000000 * USD,
 	"llama-3.1-8b-instant":                  0.05 / 1000000 * USD,
+	"llama-3.2-11b-text-preview":            0.05 / 1000000 * USD,
+	"llama-3.2-11b-vision-preview":          0.05 / 1000000 * USD,
+	"llama-3.2-1b-preview":                  0.05 / 1000000 * USD,
+	"llama-3.2-3b-preview":                  0.05 / 1000000 * USD,
+	"llama-3.2-90b-text-preview":            0.59 / 1000000 * USD,
+	"llama-guard-3-8b":                      0.05 / 1000000 * USD,
+	"llama3-70b-8192":                       0.59 / 1000000 * USD,
+	"llama3-8b-8192":                        0.05 / 1000000 * USD,
 	"llama3-groq-70b-8192-tool-use-preview": 0.89 / 1000000 * USD,
 	"llama3-groq-8b-8192-tool-use-preview":  0.19 / 1000000 * USD,
+	"mixtral-8x7b-32768":                    0.24 / 1000000 * USD,
+
 	// https://platform.lingyiwanwu.com/docs#-计费单元
 	"yi-34b-chat-0205": 2.5 / 1000 * RMB,
 	"yi-34b-chat-200k": 12.0 / 1000 * RMB,
@@ -200,6 +209,8 @@ var ModelRatio = map[string]float64{
 	"deepl-zh": 25.0 / 1000 * USD,
 	"deepl-en": 25.0 / 1000 * USD,
 	"deepl-ja": 25.0 / 1000 * USD,
+	// https://console.x.ai/
+	"grok-beta": 5.0 / 1000 * USD,
 }
 
 var CompletionRatio = map[string]float64{
@@ -364,6 +375,8 @@ func GetCompletionRatio(name string, channelType int) float64 {
 		return 3
 	case "command-r-plus":
 		return 5
+	case "grok-beta":
+		return 3
 	}
 	return 1
 }

relay/channeltype/define.go

@@ -46,5 +46,6 @@ const (
 	VertextAI
 	Proxy
 	SiliconFlow
+	XAI
 	Dummy
 )

relay/channeltype/url.go

@@ -45,7 +45,8 @@ var ChannelBaseURLs = []string{
 	"https://api.novita.ai/v3/openai",           // 41
 	"",                                          // 42
 	"",                                          // 43
-	"https://api.siliconflow.cn",                 // 44
+	"https://api.siliconflow.cn",                // 44
+	"https://api.x.ai",                          // 45
 }
 
 func init() {

relay/controller/text.go

@@ -4,6 +4,7 @@ import (
 	"bytes"
 	"encoding/json"
 	"fmt"
+	"github.com/songquanpeng/one-api/common/config"
 	"io"
 	"net/http"
 
@@ -84,7 +85,7 @@ func RelayTextHelper(c *gin.Context) *model.ErrorWithStatusCode {
 }
 
 func getRequestBody(c *gin.Context, meta *meta.Meta, textRequest *model.GeneralOpenAIRequest, adaptor adaptor.Adaptor) (io.Reader, error) {
-	if meta.APIType == apitype.OpenAI && meta.OriginModelName == meta.ActualModelName && meta.ChannelType != channeltype.Baichuan {
+	if !config.EnforceIncludeUsage && meta.APIType == apitype.OpenAI && meta.OriginModelName == meta.ActualModelName && meta.ChannelType != channeltype.Baichuan {
 		// no need to convert request for openai
 		return c.Request.Body, nil
 	}

relay/model/constant.go

@@ -1,6 +1,7 @@
 package model
 
 const (
-	ContentTypeText     = "text"
-	ContentTypeImageURL = "image_url"
+	ContentTypeText       = "text"
+	ContentTypeImageURL   = "image_url"
+	ContentTypeInputAudio = "input_audio"
 )

relay/model/general.go

@@ -12,32 +12,59 @@ type JSONSchema struct {
 	Strict      *bool                  `json:"strict,omitempty"`
 }
 
+type Audio struct {
+	Voice  string `json:"voice,omitempty"`
+	Format string `json:"format,omitempty"`
+}
+
+type StreamOptions struct {
+	IncludeUsage bool `json:"include_usage,omitempty"`
+}
+
 type GeneralOpenAIRequest struct {
-	Messages         []Message       `json:"messages,omitempty"`
-	Model            string          `json:"model,omitempty"`
-	FrequencyPenalty float64         `json:"frequency_penalty,omitempty"`
-	MaxTokens        int             `json:"max_tokens,omitempty"`
-	N                int             `json:"n,omitempty"`
-	PresencePenalty  float64         `json:"presence_penalty,omitempty"`
-	ResponseFormat   *ResponseFormat `json:"response_format,omitempty"`
-	Seed             float64         `json:"seed,omitempty"`
-	Stop             any             `json:"stop,omitempty"`
-	Stream           bool            `json:"stream,omitempty"`
-	Temperature      float64         `json:"temperature,omitempty"`
-	TopP             float64         `json:"top_p,omitempty"`
-	TopK             int             `json:"top_k,omitempty"`
-	Tools            []Tool          `json:"tools,omitempty"`
-	ToolChoice       any             `json:"tool_choice,omitempty"`
-	FunctionCall     any             `json:"function_call,omitempty"`
-	Functions        any             `json:"functions,omitempty"`
-	User             string          `json:"user,omitempty"`
-	Prompt           any             `json:"prompt,omitempty"`
-	Input            any             `json:"input,omitempty"`
-	EncodingFormat   string          `json:"encoding_format,omitempty"`
-	Dimensions       int             `json:"dimensions,omitempty"`
-	Instruction      string          `json:"instruction,omitempty"`
-	Size             string          `json:"size,omitempty"`
-	NumCtx           int         	 `json:"num_ctx,omitempty"`
+	// https://platform.openai.com/docs/api-reference/chat/create
+	Messages            []Message       `json:"messages,omitempty"`
+	Model               string          `json:"model,omitempty"`
+	Store               *bool           `json:"store,omitempty"`
+	Metadata            any             `json:"metadata,omitempty"`
+	FrequencyPenalty    *float64        `json:"frequency_penalty,omitempty"`
+	LogitBias           any             `json:"logit_bias,omitempty"`
+	Logprobs            *bool           `json:"logprobs,omitempty"`
+	TopLogprobs         *int            `json:"top_logprobs,omitempty"`
+	MaxTokens           int             `json:"max_tokens,omitempty"`
+	MaxCompletionTokens *int            `json:"max_completion_tokens,omitempty"`
+	N                   int             `json:"n,omitempty"`
+	Modalities          []string        `json:"modalities,omitempty"`
+	Prediction          any             `json:"prediction,omitempty"`
+	Audio               *Audio          `json:"audio,omitempty"`
+	PresencePenalty     *float64        `json:"presence_penalty,omitempty"`
+	ResponseFormat      *ResponseFormat `json:"response_format,omitempty"`
+	Seed                float64         `json:"seed,omitempty"`
+	ServiceTier         *string         `json:"service_tier,omitempty"`
+	Stop                any             `json:"stop,omitempty"`
+	Stream              bool            `json:"stream,omitempty"`
+	StreamOptions       *StreamOptions  `json:"stream_options,omitempty"`
+	Temperature         *float64        `json:"temperature,omitempty"`
+	TopP                *float64        `json:"top_p,omitempty"`
+	TopK                int             `json:"top_k,omitempty"`
+	Tools               []Tool          `json:"tools,omitempty"`
+	ToolChoice          any             `json:"tool_choice,omitempty"`
+	ParallelTooCalls    *bool           `json:"parallel_tool_calls,omitempty"`
+	User                string          `json:"user,omitempty"`
+	FunctionCall        any             `json:"function_call,omitempty"`
+	Functions           any             `json:"functions,omitempty"`
+	// https://platform.openai.com/docs/api-reference/embeddings/create
+	Input          any    `json:"input,omitempty"`
+	EncodingFormat string `json:"encoding_format,omitempty"`
+	Dimensions     int    `json:"dimensions,omitempty"`
+	// https://platform.openai.com/docs/api-reference/images/create
+	Prompt  any     `json:"prompt,omitempty"`
+	Quality *string `json:"quality,omitempty"`
+	Size    string  `json:"size,omitempty"`
+	Style   *string `json:"style,omitempty"`
+	// Others
+	Instruction string `json:"instruction,omitempty"`
+	NumCtx      int    `json:"num_ctx,omitempty"`
 }
 
 func (r GeneralOpenAIRequest) ParseInput() []string {

web/air/src/components/TokensTable.js

@@ -395,7 +395,7 @@ const TokensTable = () => {
         url = mjLink + `/#/?settings={"key":"sk-${key}","url":"${serverAddress}"}`;
         break;
       case 'lobechat':
-        url = chatLink + `/?settings={"keyVaults":{"openai":{"apiKey":"sk-${key}","baseURL":"${serverAddress}"/v1"}}}`;
+        url = chatLink + `/?settings={"keyVaults":{"openai":{"apiKey":"sk-${key}","baseURL":"${serverAddress}/v1"}}}`;
         break;
       default:
         if (!chatLink) {

web/air/src/constants/channel.constants.js

@@ -30,6 +30,7 @@ export const CHANNEL_OPTIONS = [
   { key: 42, text: 'VertexAI', value: 42, color: 'blue' },
   { key: 43, text: 'Proxy', value: 43, color: 'blue' },
   { key: 44, text: 'SiliconFlow', value: 44, color: 'blue' },
+  { key: 45, text: 'xAI', value: 45, color: 'blue' },
   { key: 8, text: '自定义渠道', value: 8, color: 'pink' },
   { key: 22, text: '知识库：FastGPT', value: 22, color: 'blue' },
   { key: 21, text: '知识库：AI Proxy', value: 21, color: 'purple' },

web/air/src/pages/Channel/EditChannel.js

@@ -63,7 +63,7 @@ const EditChannel = (props) => {
             let localModels = [];
             switch (value) {
                 case 14:
-                    localModels = ["claude-instant-1.2", "claude-2", "claude-2.0", "claude-2.1", "claude-3-opus-20240229", "claude-3-sonnet-20240229", "claude-3-haiku-20240307", "claude-3-5-sonnet-20240620"];
+                    localModels = ["claude-instant-1.2", "claude-2", "claude-2.0", "claude-2.1", "claude-3-opus-20240229", "claude-3-sonnet-20240229", "claude-3-haiku-20240307", "claude-3-5-haiku-20241022", "claude-3-5-sonnet-20240620", "claude-3-5-sonnet-20241022"];
                     break;
                 case 11:
                     localModels = ['PaLM-2'];
@@ -78,7 +78,7 @@ const EditChannel = (props) => {
                     localModels = ['chatglm_pro', 'chatglm_std', 'chatglm_lite'];
                     break;
                 case 18:
-                    localModels = ['SparkDesk', 'SparkDesk-v1.1', 'SparkDesk-v2.1', 'SparkDesk-v3.1', 'SparkDesk-v3.1-128K', 'SparkDesk-v3.5', 'SparkDesk-v4.0'];
+                    localModels = ['SparkDesk', 'SparkDesk-v1.1', 'SparkDesk-v2.1', 'SparkDesk-v3.1', 'SparkDesk-v3.1-128K', 'SparkDesk-v3.5', 'SparkDesk-v3.5-32K', 'SparkDesk-v4.0'];
                     break;
                 case 19:
                     localModels = ['360GPT_S2_V9', 'embedding-bert-512-v1', 'embedding_s1_v1', 'semantic_similarity_s1_v1'];

web/berry/src/constants/ChannelConstants.js

@@ -179,6 +179,12 @@ export const CHANNEL_OPTIONS = {
     value: 44,
     color: 'primary'
   },
+  45: {
+    key: 45,
+    text: 'xAI',
+    value: 45,
+    color: 'primary'
+  },
   41: {
     key: 41,
     text: 'Novita',

web/berry/src/views/Channel/type/Config.js

@@ -91,7 +91,7 @@ const typeConfig = {
       other: '版本号'
     },
     input: {
-      models: ['SparkDesk', 'SparkDesk-v1.1', 'SparkDesk-v2.1', 'SparkDesk-v3.1', 'SparkDesk-v3.1-128K', 'SparkDesk-v3.5', 'SparkDesk-v4.0']
+      models: ['SparkDesk', 'SparkDesk-v1.1', 'SparkDesk-v2.1', 'SparkDesk-v3.1', 'SparkDesk-v3.1-128K', 'SparkDesk-v3.5', 'SparkDesk-v3.5-32K', 'SparkDesk-v4.0']
     },
     prompt: {
       key: '按照如下格式输入：APPID|APISecret|APIKey',
@@ -223,6 +223,9 @@ const typeConfig = {
     },
     modelGroup: 'anthropic'
   },
+  45: {
+    modelGroup: 'xai'
+  },
 };
 
 export { defaultConfig, typeConfig };

web/berry/src/views/Token/component/TableRow.js

@@ -33,7 +33,7 @@ const COPY_OPTIONS = [
   },
   { key: 'ama', text: 'BotGem', url: 'ama://set-api-key?server={serverAddress}&key=sk-{key}', encode: true },
   { key: 'opencat', text: 'OpenCat', url: 'opencat://team/join?domain={serverAddress}&token=sk-{key}', encode: true },
-  { key: 'lobechat', text: 'LobeChat', url: 'https://lobehub.com/?settings={"keyVaults":{"openai":{"apiKey":"user-key","baseURL":"https://your-proxy.com/v1"}}}', encode: true }
+  { key: 'lobechat', text: 'LobeChat', url: 'https://lobehub.com/?settings={"keyVaults":{"openai":{"apiKey":"sk-{key}","baseURL":"{serverAddress}"}}}', encode: true }
 ];
 
 function replacePlaceholders(text, key, serverAddress) {

web/default/src/components/ChannelsTable.js

@@ -59,6 +59,12 @@ function renderBalance(type, balance) {
   }
 }
 
+function isShowDetail() {
+  return localStorage.getItem("show_detail") === "true";
+}
+
+const promptID = "detail"
+
 const ChannelsTable = () => {
   const [channels, setChannels] = useState([]);
   const [loading, setLoading] = useState(true);
@@ -66,7 +72,8 @@ const ChannelsTable = () => {
   const [searchKeyword, setSearchKeyword] = useState('');
   const [searching, setSearching] = useState(false);
   const [updatingBalance, setUpdatingBalance] = useState(false);
-  const [showPrompt, setShowPrompt] = useState(shouldShowPrompt("channel-test"));
+  const [showPrompt, setShowPrompt] = useState(shouldShowPrompt(promptID));
+  const [showDetail, setShowDetail] = useState(isShowDetail());
 
   const loadChannels = async (startIdx) => {
     const res = await API.get(`/api/channel/?p=${startIdx}`);
@@ -120,6 +127,11 @@ const ChannelsTable = () => {
     await loadChannels(activePage - 1);
   };
 
+  const toggleShowDetail = () => {
+    setShowDetail(!showDetail);
+    localStorage.setItem("show_detail", (!showDetail).toString());
+  }
+
   useEffect(() => {
     loadChannels(0)
       .then()
@@ -364,11 +376,13 @@ const ChannelsTable = () => {
         showPrompt && (
           <Message onDismiss={() => {
             setShowPrompt(false);
-            setPromptShown("channel-test");
+            setPromptShown(promptID);
           }}>
             OpenAI 渠道已经不再支持通过 key 获取余额，因此余额显示为 0。对于支持的渠道类型，请点击余额进行刷新。
             <br/>
             渠道测试仅支持 chat 模型，优先使用 gpt-3.5-turbo，如果该模型不可用则使用你所配置的模型列表中的第一个模型。
+            <br/>
+            点击下方详情按钮可以显示余额以及设置额外的测试模型。
           </Message>
         )
       }
@@ -428,6 +442,7 @@ const ChannelsTable = () => {
               onClick={() => {
                 sortChannel('balance');
               }}
+              hidden={!showDetail}
             >
               余额
             </Table.HeaderCell>
@@ -439,7 +454,7 @@ const ChannelsTable = () => {
             >
               优先级
             </Table.HeaderCell>
-            <Table.HeaderCell>测试模型</Table.HeaderCell>
+            <Table.HeaderCell hidden={!showDetail}>测试模型</Table.HeaderCell>
             <Table.HeaderCell>操作</Table.HeaderCell>
           </Table.Row>
         </Table.Header>
@@ -467,7 +482,7 @@ const ChannelsTable = () => {
                       basic
                     />
                   </Table.Cell>
-                  <Table.Cell>
+                  <Table.Cell hidden={!showDetail}>
                     <Popup
                       trigger={<span onClick={() => {
                         updateChannelBalance(channel.id, channel.name, idx);
@@ -494,7 +509,7 @@ const ChannelsTable = () => {
                       basic
                     />
                   </Table.Cell>
-                  <Table.Cell>
+                  <Table.Cell hidden={!showDetail}>
                     <Dropdown
                       placeholder='请选择测试模型'
                       selection
@@ -573,7 +588,7 @@ const ChannelsTable = () => {
 
         <Table.Footer>
           <Table.Row>
-            <Table.HeaderCell colSpan='9'>
+            <Table.HeaderCell colSpan={showDetail ? "10" : "8"}>
               <Button size='small' as={Link} to='/channel/add' loading={loading}>
                 添加新的渠道
               </Button>
@@ -611,6 +626,7 @@ const ChannelsTable = () => {
                 }
               />
               <Button size='small' onClick={refresh} loading={loading}>刷新</Button>
+              <Button size='small' onClick={toggleShowDetail}>{showDetail ? "隐藏详情" : "详情"}</Button>
             </Table.HeaderCell>
           </Table.Row>
         </Table.Footer>

web/default/src/components/TokensTable.js

@@ -117,7 +117,7 @@ const TokensTable = () => {
         url = nextUrl;
         break;
       case 'lobechat':
-        url = nextLink + `/?settings={"keyVaults":{"openai":{"apiKey":"sk-${key}","baseURL":"${serverAddress}"/v1"}}}`;
+        url = nextLink + `/?settings={"keyVaults":{"openai":{"apiKey":"sk-${key}","baseURL":"${serverAddress}/v1"}}}`;
         break;
       default:
         url = `sk-${key}`;
@@ -160,7 +160,7 @@ const TokensTable = () => {
         break;
 
       case 'lobechat':
-        url = chatLink + `/?settings={"keyVaults":{"openai":{"apiKey":"sk-${key}","baseURL":"${serverAddress}"/v1"}}}`;
+        url = chatLink + `/?settings={"keyVaults":{"openai":{"apiKey":"sk-${key}","baseURL":"${serverAddress}/v1"}}}`;
         break;
 
       default:

web/default/src/constants/channel.constants.js

@@ -30,6 +30,7 @@ export const CHANNEL_OPTIONS = [
     { key: 42, text: 'VertexAI', value: 42, color: 'blue' },
     { key: 43, text: 'Proxy', value: 43, color: 'blue' },
     { key: 44, text: 'SiliconFlow', value: 44, color: 'blue' },
+    { key: 45, text: 'xAI', value: 45, color: 'blue' },
     { key: 8, text: '自定义渠道', value: 8, color: 'pink' },
     { key: 22, text: '知识库：FastGPT', value: 22, color: 'blue' },
     { key: 21, text: '知识库：AI Proxy', value: 21, color: 'purple' },

web/default/src/pages/Channel/index.js

@@ -2,7 +2,7 @@ import React from 'react';
 import { Header, Segment } from 'semantic-ui-react';
 import ChannelsTable from '../../components/ChannelsTable';
 
-const File = () => (
+const Channel = () => (
   <>
     <Segment>
       <Header as='h3'>管理渠道</Header>
@@ -11,4 +11,4 @@ const File = () => (
   </>
 );
 
-export default File;
+export default Channel;