xinyin025/one-api
1
0
Fork 0
mirror of https://github.com/songquanpeng/one-api.git synced 2025-11-09 10:13:42 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity

feat: 合并main

Browse Source
This commit is contained in:
suziheng
2025-04-21 16:36:54 +08:00
parent dfcf8868fe 8df4a2670b
commit 77267aa1b8
138 changed files with 8404 additions and 3879 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
relay/adaptor/ali/constants.go
View File

@@ -14,10 +14,14 @@ var ModelList = []string{
"qwen2-72b-instruct", "qwen2-57b-a14b-instruct", "qwen2-7b-instruct", "qwen2-1.5b-instruct", "qwen2-0.5b-instruct",
"qwen1.5-110b-chat", "qwen1.5-72b-chat", "qwen1.5-32b-chat", "qwen1.5-14b-chat", "qwen1.5-7b-chat", "qwen1.5-1.8b-chat", "qwen1.5-0.5b-chat",
"qwen-72b-chat", "qwen-14b-chat", "qwen-7b-chat", "qwen-1.8b-chat", "qwen-1.8b-longcontext-chat",
"qvq-72b-preview",
"qwen2.5-vl-72b-instruct", "qwen2.5-vl-7b-instruct", "qwen2.5-vl-2b-instruct", "qwen2.5-vl-1b-instruct", "qwen2.5-vl-0.5b-instruct",
"qwen2-vl-7b-instruct", "qwen2-vl-2b-instruct", "qwen-vl-v1", "qwen-vl-chat-v1",
"qwen2-audio-instruct", "qwen-audio-chat",
"qwen2.5-math-72b-instruct", "qwen2.5-math-7b-instruct", "qwen2.5-math-1.5b-instruct", "qwen2-math-72b-instruct", "qwen2-math-7b-instruct", "qwen2-math-1.5b-instruct",
"qwen2.5-coder-32b-instruct", "qwen2.5-coder-14b-instruct", "qwen2.5-coder-7b-instruct", "qwen2.5-coder-3b-instruct", "qwen2.5-coder-1.5b-instruct", "qwen2.5-coder-0.5b-instruct",
"text-embedding-v1", "text-embedding-v3", "text-embedding-v2", "text-embedding-async-v2", "text-embedding-async-v1",
"ali-stable-diffusion-xl", "ali-stable-diffusion-v1.5", "wanx-v1",
"qwen-mt-plus", "qwen-mt-turbo",
"deepseek-r1", "deepseek-v3", "deepseek-r1-distill-qwen-1.5b", "deepseek-r1-distill-qwen-7b", "deepseek-r1-distill-qwen-14b", "deepseek-r1-distill-qwen-32b", "deepseek-r1-distill-llama-8b", "deepseek-r1-distill-llama-70b",
}

20
relay/adaptor/alibailian/constants.go Normal file
View File

@@ -0,0 +1,20 @@
package alibailian
// https://help.aliyun.com/zh/model-studio/getting-started/models
var ModelList = []string{
"qwen-turbo",
"qwen-plus",
"qwen-long",
"qwen-max",
"qwen-coder-plus",
"qwen-coder-plus-latest",
"qwen-coder-turbo",
"qwen-coder-turbo-latest",
"qwen-mt-plus",
"qwen-mt-turbo",
"qwq-32b-preview",
"deepseek-r1",
"deepseek-v3",
}

19
relay/adaptor/alibailian/main.go Normal file
View File

@@ -0,0 +1,19 @@
package alibailian
import (
"fmt"
"github.com/songquanpeng/one-api/relay/meta"
"github.com/songquanpeng/one-api/relay/relaymode"
)
func GetRequestURL(meta *meta.Meta) (string, error) {
switch meta.Mode {
case relaymode.ChatCompletions:
return fmt.Sprintf("%s/compatible-mode/v1/chat/completions", meta.BaseURL), nil
case relaymode.Embeddings:
return fmt.Sprintf("%s/compatible-mode/v1/embeddings", meta.BaseURL), nil
default:
}
return "", fmt.Errorf("unsupported relay mode %d for ali bailian", meta.Mode)
}

1
relay/adaptor/anthropic/constants.go
View File

@@ -4,6 +4,7 @@ var ModelList = []string{
"claude-instant-1.2", "claude-2.0", "claude-2.1",
"claude-3-haiku-20240307",
"claude-3-5-haiku-20241022",
"claude-3-5-haiku-latest",
"claude-3-sonnet-20240229",
"claude-3-opus-20240229",
"claude-3-5-sonnet-20240620",

30
relay/adaptor/baiduv2/constants.go Normal file
View File

@@ -0,0 +1,30 @@
package baiduv2
// https://console.bce.baidu.com/support/?_=1692863460488&timestamp=1739074632076#/api?product=QIANFAN&project=%E5%8D%83%E5%B8%86ModelBuilder&parent=%E5%AF%B9%E8%AF%9DChat%20V2&api=v2%2Fchat%2Fcompletions&method=post
// https://cloud.baidu.com/doc/WENXINWORKSHOP/s/Fm2vrveyu#%E6%94%AF%E6%8C%81%E6%A8%A1%E5%9E%8B%E5%88%97%E8%A1%A8
var ModelList = []string{
"ernie-4.0-8k-latest",
"ernie-4.0-8k-preview",
"ernie-4.0-8k",
"ernie-4.0-turbo-8k-latest",
"ernie-4.0-turbo-8k-preview",
"ernie-4.0-turbo-8k",
"ernie-4.0-turbo-128k",
"ernie-3.5-8k-preview",
"ernie-3.5-8k",
"ernie-3.5-128k",
"ernie-speed-8k",
"ernie-speed-128k",
"ernie-speed-pro-128k",
"ernie-lite-8k",
"ernie-lite-pro-128k",
"ernie-tiny-8k",
"ernie-char-8k",
"ernie-char-fiction-8k",
"ernie-novel-8k",
"deepseek-v3",
"deepseek-r1",
"deepseek-r1-distill-qwen-32b",
"deepseek-r1-distill-qwen-14b",
}

17
relay/adaptor/baiduv2/main.go Normal file
View File

@@ -0,0 +1,17 @@
package baiduv2
import (
"fmt"
"github.com/songquanpeng/one-api/relay/meta"
"github.com/songquanpeng/one-api/relay/relaymode"
)
func GetRequestURL(meta *meta.Meta) (string, error) {
switch meta.Mode {
case relaymode.ChatCompletions:
return fmt.Sprintf("%s/v2/chat/completions", meta.BaseURL), nil
default:
}
return "", fmt.Errorf("unsupported relay mode %d for baidu v2", meta.Mode)
}

2
relay/adaptor/deepseek/constants.go
View File

@@ -2,5 +2,5 @@ package deepseek
var ModelList = []string{
"deepseek-chat",
"deepseek-coder",
"deepseek-reasoner",
}

5
relay/adaptor/gemini/adaptor.go
View File

@@ -5,6 +5,7 @@ import (
"fmt"
"io"
"net/http"
"strings"
"github.com/gin-gonic/gin"
"github.com/songquanpeng/one-api/common/config"
@@ -20,12 +21,12 @@ type Adaptor struct {
}
func (a *Adaptor) Init(meta *meta.Meta) {
}
func (a *Adaptor) GetRequestURL(meta *meta.Meta) (string, error) {
defaultVersion := config.GeminiVersion
if meta.ActualModelName == "gemini-2.0-flash-exp" {
if strings.Contains(meta.ActualModelName, "gemini-2.0") ||
strings.Contains(meta.ActualModelName, "gemini-1.5") {
defaultVersion = "v1beta"
}

36
relay/adaptor/gemini/constants.go
View File

@@ -1,11 +1,35 @@
package gemini
import (
"github.com/songquanpeng/one-api/relay/adaptor/geminiv2"
)
// https://ai.google.dev/models/gemini
var ModelList = []string{
"gemini-pro", "gemini-1.0-pro",
"gemini-1.5-flash", "gemini-1.5-pro",
"text-embedding-004", "aqa",
"gemini-2.0-flash-exp",
"gemini-2.0-flash-thinking-exp",
var ModelList = geminiv2.ModelList
// ModelsSupportSystemInstruction is the list of models that support system instruction.
//
// https://cloud.google.com/vertex-ai/generative-ai/docs/learn/prompts/system-instructions
var ModelsSupportSystemInstruction = []string{
// "gemini-1.0-pro-002",
// "gemini-1.5-flash", "gemini-1.5-flash-001", "gemini-1.5-flash-002",
// "gemini-1.5-flash-8b",
// "gemini-1.5-pro", "gemini-1.5-pro-001", "gemini-1.5-pro-002",
// "gemini-1.5-pro-experimental",
"gemini-2.0-flash", "gemini-2.0-flash-exp",
"gemini-2.0-flash-thinking-exp-01-21",
}
// IsModelSupportSystemInstruction check if the model support system instruction.
//
// Because the main version of Go is 1.20, slice.Contains cannot be used
func IsModelSupportSystemInstruction(model string) bool {
for _, m := range ModelsSupportSystemInstruction {
if m == model {
return true
}
}
return false
}

9
relay/adaptor/gemini/main.go
View File

@@ -132,9 +132,16 @@ func ConvertRequest(textRequest model.GeneralOpenAIRequest) *ChatRequest {
}
// Converting system prompt to prompt from user for the same reason
if content.Role == "system" {
content.Role = "user"
shouldAddDummyModelMessage = true
if IsModelSupportSystemInstruction(textRequest.Model) {
geminiRequest.SystemInstruction = &content
geminiRequest.SystemInstruction.Role = ""
continue
} else {
content.Role = "user"
}
}
geminiRequest.Contents = append(geminiRequest.Contents, content)
// If a system message is the last message, we need to add a dummy model message to make gemini happy

9
relay/adaptor/gemini/model.go
View File

@@ -1,10 +1,11 @@
package gemini
type ChatRequest struct {
Contents []ChatContent `json:"contents"`
SafetySettings []ChatSafetySettings `json:"safety_settings,omitempty"`
GenerationConfig ChatGenerationConfig `json:"generation_config,omitempty"`
Tools []ChatTools `json:"tools,omitempty"`
Contents []ChatContent `json:"contents"`
SafetySettings []ChatSafetySettings `json:"safety_settings,omitempty"`
GenerationConfig ChatGenerationConfig `json:"generation_config,omitempty"`
Tools []ChatTools `json:"tools,omitempty"`
SystemInstruction *ChatContent `json:"system_instruction,omitempty"`
}
type EmbeddingRequest struct {

15
relay/adaptor/geminiv2/constants.go Normal file
View File

@@ -0,0 +1,15 @@
package geminiv2
// https://ai.google.dev/models/gemini
var ModelList = []string{
"gemini-pro", "gemini-1.0-pro",
// "gemma-2-2b-it", "gemma-2-9b-it", "gemma-2-27b-it",
"gemini-1.5-flash", "gemini-1.5-flash-8b",
"gemini-1.5-pro", "gemini-1.5-pro-experimental",
"text-embedding-004", "aqa",
"gemini-2.0-flash", "gemini-2.0-flash-exp",
"gemini-2.0-flash-lite-preview-02-05",
"gemini-2.0-flash-thinking-exp-01-21",
"gemini-2.0-pro-exp-02-05",
}

14
relay/adaptor/geminiv2/main.go Normal file
View File

@@ -0,0 +1,14 @@
package geminiv2
import (
"fmt"
"strings"
"github.com/songquanpeng/one-api/relay/meta"
)
func GetRequestURL(meta *meta.Meta) (string, error) {
baseURL := strings.TrimSuffix(meta.BaseURL, "/")
requestPath := strings.TrimPrefix(meta.RequestURLPath, "/v1")
return fmt.Sprintf("%s%s", baseURL, requestPath), nil
}

4
relay/adaptor/groq/constants.go
View File

@@ -3,7 +3,6 @@ package groq
// https://console.groq.com/docs/models
var ModelList = []string{
"gemma-7b-it",
"gemma2-9b-it",
"llama-3.1-70b-versatile",
"llama-3.1-8b-instant",
@@ -11,7 +10,6 @@ var ModelList = []string{
"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",
@@ -24,4 +22,6 @@ var ModelList = []string{
"distil-whisper-large-v3-en",
"whisper-large-v3",
"whisper-large-v3-turbo",
"deepseek-r1-distill-llama-70b-specdec",
"deepseek-r1-distill-llama-70b",
}

2
relay/adaptor/minimax/constants.go
View File

@@ -8,4 +8,6 @@ var ModelList = []string{
"abab6-chat",
"abab5.5-chat",
"abab5.5s-chat",
"MiniMax-VL-01",
"MiniMax-Text-01",
}

10
relay/adaptor/openai/adaptor.go
View File

@@ -8,8 +8,12 @@ import (
"strings"
"github.com/gin-gonic/gin"
"github.com/songquanpeng/one-api/relay/adaptor"
"github.com/songquanpeng/one-api/relay/adaptor/alibailian"
"github.com/songquanpeng/one-api/relay/adaptor/baiduv2"
"github.com/songquanpeng/one-api/relay/adaptor/doubao"
"github.com/songquanpeng/one-api/relay/adaptor/geminiv2"
"github.com/songquanpeng/one-api/relay/adaptor/minimax"
"github.com/songquanpeng/one-api/relay/adaptor/novita"
"github.com/songquanpeng/one-api/relay/channeltype"
@@ -52,6 +56,12 @@ func (a *Adaptor) GetRequestURL(meta *meta.Meta) (string, error) {
return doubao.GetRequestURL(meta)
case channeltype.Novita:
return novita.GetRequestURL(meta)
case channeltype.BaiduV2:
return baiduv2.GetRequestURL(meta)
case channeltype.AliBailian:
return alibailian.GetRequestURL(meta)
case channeltype.GeminiOpenAICompatible:
return geminiv2.GetRequestURL(meta)
default:
return GetFullRequestURL(meta.BaseURL, meta.RequestURLPath, meta.ChannelType), nil
}

17
relay/adaptor/openai/compatible.go
View File

@@ -2,19 +2,24 @@ package openai
import (
"github.com/songquanpeng/one-api/relay/adaptor/ai360"
"github.com/songquanpeng/one-api/relay/adaptor/alibailian"
"github.com/songquanpeng/one-api/relay/adaptor/baichuan"
"github.com/songquanpeng/one-api/relay/adaptor/baiduv2"
"github.com/songquanpeng/one-api/relay/adaptor/deepseek"
"github.com/songquanpeng/one-api/relay/adaptor/doubao"
"github.com/songquanpeng/one-api/relay/adaptor/geminiv2"
"github.com/songquanpeng/one-api/relay/adaptor/groq"
"github.com/songquanpeng/one-api/relay/adaptor/lingyiwanwu"
"github.com/songquanpeng/one-api/relay/adaptor/minimax"
"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/openrouter"
"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/xai"
"github.com/songquanpeng/one-api/relay/adaptor/xunfeiv2"
"github.com/songquanpeng/one-api/relay/channeltype"
)
@@ -34,6 +39,8 @@ var CompatibleChannels = []int{
channeltype.Novita,
channeltype.SiliconFlow,
channeltype.XAI,
channeltype.BaiduV2,
channeltype.XunfeiV2,
}
func GetCompatibleChannelMeta(channelType int) (string, []string) {
@@ -68,6 +75,16 @@ func GetCompatibleChannelMeta(channelType int) (string, []string) {
return "siliconflow", siliconflow.ModelList
case channeltype.XAI:
return "xai", xai.ModelList
case channeltype.BaiduV2:
return "baiduv2", baiduv2.ModelList
case channeltype.XunfeiV2:
return "xunfeiv2", xunfeiv2.ModelList
case channeltype.OpenRouter:
return "openrouter", openrouter.ModelList
case channeltype.AliBailian:
return "alibailian", alibailian.ModelList
case channeltype.GeminiOpenAICompatible:
return "geminiv2", geminiv2.ModelList
default:
return "openai", ModelList
}

3
relay/adaptor/openai/helper.go
View File

@@ -17,6 +17,9 @@ func ResponseText2Usage(responseText string, modelName string, promptTokens int)
}
func GetFullRequestURL(baseURL string, requestURL string, channelType int) string {
if channelType == channeltype.OpenAICompatible {
return fmt.Sprintf("%s%s", strings.TrimSuffix(baseURL, "/"), strings.TrimPrefix(requestURL, "/v1"))
}
fullRequestURL := fmt.Sprintf("%s%s", baseURL, requestURL)
if strings.HasPrefix(baseURL, "https://gateway.ai.cloudflare.com") {

9
relay/adaptor/openai/token.go
View File

@@ -3,14 +3,16 @@ package openai
import (
"errors"
"fmt"
"math"
"strings"
"github.com/pkoukk/tiktoken-go"
"github.com/songquanpeng/one-api/common/config"
"github.com/songquanpeng/one-api/common/image"
"github.com/songquanpeng/one-api/common/logger"
billingratio "github.com/songquanpeng/one-api/relay/billing/ratio"
"github.com/songquanpeng/one-api/relay/model"
"math"
"strings"
)
// tokenEncoderMap won't grow after initialization
@@ -21,7 +23,8 @@ func InitTokenEncoders() {
logger.SysLog("initializing token encoders")
gpt35TokenEncoder, err := tiktoken.EncodingForModel("gpt-3.5-turbo")
if err != nil {
logger.FatalLog(fmt.Sprintf("failed to get gpt-3.5-turbo token encoder: %s", err.Error()))
logger.FatalLog(fmt.Sprintf("failed to get gpt-3.5-turbo token encoder: %s, "+
"if you are using in offline environment, please set TIKTOKEN_CACHE_DIR to use exsited files, check this link for more information: https://stackoverflow.com/questions/76106366/how-to-use-tiktoken-in-offline-mode-computer ", err.Error()))
}
defaultTokenEncoder = gpt35TokenEncoder
gpt4oTokenEncoder, err := tiktoken.EncodingForModel("gpt-4o")

235
relay/adaptor/openrouter/constants.go Normal file
View File

@@ -0,0 +1,235 @@
package openrouter
var ModelList = []string{
"01-ai/yi-large",
"aetherwiing/mn-starcannon-12b",
"ai21/jamba-1-5-large",
"ai21/jamba-1-5-mini",
"ai21/jamba-instruct",
"aion-labs/aion-1.0",
"aion-labs/aion-1.0-mini",
"aion-labs/aion-rp-llama-3.1-8b",
"allenai/llama-3.1-tulu-3-405b",
"alpindale/goliath-120b",
"alpindale/magnum-72b",
"amazon/nova-lite-v1",
"amazon/nova-micro-v1",
"amazon/nova-pro-v1",
"anthracite-org/magnum-v2-72b",
"anthracite-org/magnum-v4-72b",
"anthropic/claude-2",
"anthropic/claude-2.0",
"anthropic/claude-2.0:beta",
"anthropic/claude-2.1",
"anthropic/claude-2.1:beta",
"anthropic/claude-2:beta",
"anthropic/claude-3-haiku",
"anthropic/claude-3-haiku:beta",
"anthropic/claude-3-opus",
"anthropic/claude-3-opus:beta",
"anthropic/claude-3-sonnet",
"anthropic/claude-3-sonnet:beta",
"anthropic/claude-3.5-haiku",
"anthropic/claude-3.5-haiku-20241022",
"anthropic/claude-3.5-haiku-20241022:beta",
"anthropic/claude-3.5-haiku:beta",
"anthropic/claude-3.5-sonnet",
"anthropic/claude-3.5-sonnet-20240620",
"anthropic/claude-3.5-sonnet-20240620:beta",
"anthropic/claude-3.5-sonnet:beta",
"cognitivecomputations/dolphin-mixtral-8x22b",
"cognitivecomputations/dolphin-mixtral-8x7b",
"cohere/command",
"cohere/command-r",
"cohere/command-r-03-2024",
"cohere/command-r-08-2024",
"cohere/command-r-plus",
"cohere/command-r-plus-04-2024",
"cohere/command-r-plus-08-2024",
"cohere/command-r7b-12-2024",
"databricks/dbrx-instruct",
"deepseek/deepseek-chat",
"deepseek/deepseek-chat-v2.5",
"deepseek/deepseek-chat:free",
"deepseek/deepseek-r1",
"deepseek/deepseek-r1-distill-llama-70b",
"deepseek/deepseek-r1-distill-llama-70b:free",
"deepseek/deepseek-r1-distill-llama-8b",
"deepseek/deepseek-r1-distill-qwen-1.5b",
"deepseek/deepseek-r1-distill-qwen-14b",
"deepseek/deepseek-r1-distill-qwen-32b",
"deepseek/deepseek-r1:free",
"eva-unit-01/eva-llama-3.33-70b",
"eva-unit-01/eva-qwen-2.5-32b",
"eva-unit-01/eva-qwen-2.5-72b",
"google/gemini-2.0-flash-001",
"google/gemini-2.0-flash-exp:free",
"google/gemini-2.0-flash-lite-preview-02-05:free",
"google/gemini-2.0-flash-thinking-exp-1219:free",
"google/gemini-2.0-flash-thinking-exp:free",
"google/gemini-2.0-pro-exp-02-05:free",
"google/gemini-exp-1206:free",
"google/gemini-flash-1.5",
"google/gemini-flash-1.5-8b",
"google/gemini-flash-1.5-8b-exp",
"google/gemini-pro",
"google/gemini-pro-1.5",
"google/gemini-pro-vision",
"google/gemma-2-27b-it",
"google/gemma-2-9b-it",
"google/gemma-2-9b-it:free",
"google/gemma-7b-it",
"google/learnlm-1.5-pro-experimental:free",
"google/palm-2-chat-bison",
"google/palm-2-chat-bison-32k",
"google/palm-2-codechat-bison",
"google/palm-2-codechat-bison-32k",
"gryphe/mythomax-l2-13b",
"gryphe/mythomax-l2-13b:free",
"huggingfaceh4/zephyr-7b-beta:free",
"infermatic/mn-inferor-12b",
"inflection/inflection-3-pi",
"inflection/inflection-3-productivity",
"jondurbin/airoboros-l2-70b",
"liquid/lfm-3b",
"liquid/lfm-40b",
"liquid/lfm-7b",
"mancer/weaver",
"meta-llama/llama-2-13b-chat",
"meta-llama/llama-2-70b-chat",
"meta-llama/llama-3-70b-instruct",
"meta-llama/llama-3-8b-instruct",
"meta-llama/llama-3-8b-instruct:free",
"meta-llama/llama-3.1-405b",
"meta-llama/llama-3.1-405b-instruct",
"meta-llama/llama-3.1-70b-instruct",
"meta-llama/llama-3.1-8b-instruct",
"meta-llama/llama-3.2-11b-vision-instruct",
"meta-llama/llama-3.2-11b-vision-instruct:free",
"meta-llama/llama-3.2-1b-instruct",
"meta-llama/llama-3.2-3b-instruct",
"meta-llama/llama-3.2-90b-vision-instruct",
"meta-llama/llama-3.3-70b-instruct",
"meta-llama/llama-3.3-70b-instruct:free",
"meta-llama/llama-guard-2-8b",
"microsoft/phi-3-medium-128k-instruct",
"microsoft/phi-3-medium-128k-instruct:free",
"microsoft/phi-3-mini-128k-instruct",
"microsoft/phi-3-mini-128k-instruct:free",
"microsoft/phi-3.5-mini-128k-instruct",
"microsoft/phi-4",
"microsoft/wizardlm-2-7b",
"microsoft/wizardlm-2-8x22b",
"minimax/minimax-01",
"mistralai/codestral-2501",
"mistralai/codestral-mamba",
"mistralai/ministral-3b",
"mistralai/ministral-8b",
"mistralai/mistral-7b-instruct",
"mistralai/mistral-7b-instruct-v0.1",
"mistralai/mistral-7b-instruct-v0.3",
"mistralai/mistral-7b-instruct:free",
"mistralai/mistral-large",
"mistralai/mistral-large-2407",
"mistralai/mistral-large-2411",
"mistralai/mistral-medium",
"mistralai/mistral-nemo",
"mistralai/mistral-nemo:free",
"mistralai/mistral-small",
"mistralai/mistral-small-24b-instruct-2501",
"mistralai/mistral-small-24b-instruct-2501:free",
"mistralai/mistral-tiny",
"mistralai/mixtral-8x22b-instruct",
"mistralai/mixtral-8x7b",
"mistralai/mixtral-8x7b-instruct",
"mistralai/pixtral-12b",
"mistralai/pixtral-large-2411",
"neversleep/llama-3-lumimaid-70b",
"neversleep/llama-3-lumimaid-8b",
"neversleep/llama-3-lumimaid-8b:extended",
"neversleep/llama-3.1-lumimaid-70b",
"neversleep/llama-3.1-lumimaid-8b",
"neversleep/noromaid-20b",
"nothingiisreal/mn-celeste-12b",
"nousresearch/hermes-2-pro-llama-3-8b",
"nousresearch/hermes-3-llama-3.1-405b",
"nousresearch/hermes-3-llama-3.1-70b",
"nousresearch/nous-hermes-2-mixtral-8x7b-dpo",
"nousresearch/nous-hermes-llama2-13b",
"nvidia/llama-3.1-nemotron-70b-instruct",
"nvidia/llama-3.1-nemotron-70b-instruct:free",
"openai/chatgpt-4o-latest",
"openai/gpt-3.5-turbo",
"openai/gpt-3.5-turbo-0125",
"openai/gpt-3.5-turbo-0613",
"openai/gpt-3.5-turbo-1106",
"openai/gpt-3.5-turbo-16k",
"openai/gpt-3.5-turbo-instruct",
"openai/gpt-4",
"openai/gpt-4-0314",
"openai/gpt-4-1106-preview",
"openai/gpt-4-32k",
"openai/gpt-4-32k-0314",
"openai/gpt-4-turbo",
"openai/gpt-4-turbo-preview",
"openai/gpt-4o",
"openai/gpt-4o-2024-05-13",
"openai/gpt-4o-2024-08-06",
"openai/gpt-4o-2024-11-20",
"openai/gpt-4o-mini",
"openai/gpt-4o-mini-2024-07-18",
"openai/gpt-4o:extended",
"openai/o1",
"openai/o1-mini",
"openai/o1-mini-2024-09-12",
"openai/o1-preview",
"openai/o1-preview-2024-09-12",
"openai/o3-mini",
"openai/o3-mini-high",
"openchat/openchat-7b",
"openchat/openchat-7b:free",
"openrouter/auto",
"perplexity/llama-3.1-sonar-huge-128k-online",
"perplexity/llama-3.1-sonar-large-128k-chat",
"perplexity/llama-3.1-sonar-large-128k-online",
"perplexity/llama-3.1-sonar-small-128k-chat",
"perplexity/llama-3.1-sonar-small-128k-online",
"perplexity/sonar",
"perplexity/sonar-reasoning",
"pygmalionai/mythalion-13b",
"qwen/qvq-72b-preview",
"qwen/qwen-2-72b-instruct",
"qwen/qwen-2-7b-instruct",
"qwen/qwen-2-7b-instruct:free",
"qwen/qwen-2-vl-72b-instruct",
"qwen/qwen-2-vl-7b-instruct",
"qwen/qwen-2.5-72b-instruct",
"qwen/qwen-2.5-7b-instruct",
"qwen/qwen-2.5-coder-32b-instruct",
"qwen/qwen-max",
"qwen/qwen-plus",
"qwen/qwen-turbo",
"qwen/qwen-vl-plus:free",
"qwen/qwen2.5-vl-72b-instruct:free",
"qwen/qwq-32b-preview",
"raifle/sorcererlm-8x22b",
"sao10k/fimbulvetr-11b-v2",
"sao10k/l3-euryale-70b",
"sao10k/l3-lunaris-8b",
"sao10k/l3.1-70b-hanami-x1",
"sao10k/l3.1-euryale-70b",
"sao10k/l3.3-euryale-70b",
"sophosympatheia/midnight-rose-70b",
"sophosympatheia/rogue-rose-103b-v0.2:free",
"teknium/openhermes-2.5-mistral-7b",
"thedrummer/rocinante-12b",
"thedrummer/unslopnemo-12b",
"undi95/remm-slerp-l2-13b",
"undi95/toppy-m-7b",
"undi95/toppy-m-7b:free",
"x-ai/grok-2-1212",
"x-ai/grok-2-vision-1212",
"x-ai/grok-beta",
"x-ai/grok-vision-beta",
"xwin-lm/xwin-lm-70b",
}

32
relay/adaptor/tencent/adaptor.go
View File

@@ -2,16 +2,19 @@ package tencent
import (
"errors"
"io"
"net/http"
"strconv"
"strings"
"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"
"io"
"net/http"
"strconv"
"strings"
"github.com/songquanpeng/one-api/relay/relaymode"
)
// https://cloud.tencent.com/document/api/1729/101837
@@ -52,10 +55,18 @@ func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.G
if err != nil {
return nil, err
}
tencentRequest := ConvertRequest(*request)
var convertedRequest any
switch relayMode {
case relaymode.Embeddings:
a.Action = "GetEmbedding"
convertedRequest = ConvertEmbeddingRequest(*request)
default:
a.Action = "ChatCompletions"
convertedRequest = ConvertRequest(*request)
}
// we have to calculate the sign here
a.Sign = GetSign(*tencentRequest, a, secretId, secretKey)
return tencentRequest, nil
a.Sign = GetSign(convertedRequest, a, secretId, secretKey)
return convertedRequest, nil
}
func (a *Adaptor) ConvertImageRequest(request *model.ImageRequest) (any, error) {
@@ -75,7 +86,12 @@ func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Met
err, responseText = StreamHandler(c, resp)
usage = openai.ResponseText2Usage(responseText, meta.ActualModelName, meta.PromptTokens)
} else {
err, usage = Handler(c, resp)
switch meta.Mode {
case relaymode.Embeddings:
err, usage = EmbeddingHandler(c, resp)
default:
err, usage = Handler(c, resp)
}
}
return
}

1
relay/adaptor/tencent/constants.go
View File

@@ -6,4 +6,5 @@ var ModelList = []string{
"hunyuan-standard-256K",
"hunyuan-pro",
"hunyuan-vision",
"hunyuan-embedding",
}

68
relay/adaptor/tencent/main.go
View File

@@ -8,7 +8,6 @@ import (
"encoding/json"
"errors"
"fmt"
"github.com/songquanpeng/one-api/common/render"
"io"
"net/http"
"strconv"
@@ -16,11 +15,14 @@ import (
"time"
"github.com/gin-gonic/gin"
"github.com/songquanpeng/one-api/common"
"github.com/songquanpeng/one-api/common/conv"
"github.com/songquanpeng/one-api/common/ctxkey"
"github.com/songquanpeng/one-api/common/helper"
"github.com/songquanpeng/one-api/common/logger"
"github.com/songquanpeng/one-api/common/random"
"github.com/songquanpeng/one-api/common/render"
"github.com/songquanpeng/one-api/relay/adaptor/openai"
"github.com/songquanpeng/one-api/relay/constant"
"github.com/songquanpeng/one-api/relay/model"
@@ -44,8 +46,68 @@ func ConvertRequest(request model.GeneralOpenAIRequest) *ChatRequest {
}
}
func ConvertEmbeddingRequest(request model.GeneralOpenAIRequest) *EmbeddingRequest {
return &EmbeddingRequest{
InputList: request.ParseInput(),
}
}
func EmbeddingHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, *model.Usage) {
var tencentResponseP EmbeddingResponseP
err := json.NewDecoder(resp.Body).Decode(&tencentResponseP)
if err != nil {
return openai.ErrorWrapper(err, "unmarshal_response_body_failed", http.StatusInternalServerError), nil
}
err = resp.Body.Close()
if err != nil {
return openai.ErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), nil
}
tencentResponse := tencentResponseP.Response
if tencentResponse.Error.Code != "" {
return &model.ErrorWithStatusCode{
Error: model.Error{
Message: tencentResponse.Error.Message,
Code: tencentResponse.Error.Code,
},
StatusCode: resp.StatusCode,
}, nil
}
requestModel := c.GetString(ctxkey.RequestModel)
fullTextResponse := embeddingResponseTencent2OpenAI(&tencentResponse)
fullTextResponse.Model = requestModel
jsonResponse, err := json.Marshal(fullTextResponse)
if err != nil {
return openai.ErrorWrapper(err, "marshal_response_body_failed", http.StatusInternalServerError), nil
}
c.Writer.Header().Set("Content-Type", "application/json")
c.Writer.WriteHeader(resp.StatusCode)
_, err = c.Writer.Write(jsonResponse)
return nil, &fullTextResponse.Usage
}
func embeddingResponseTencent2OpenAI(response *EmbeddingResponse) *openai.EmbeddingResponse {
openAIEmbeddingResponse := openai.EmbeddingResponse{
Object: "list",
Data: make([]openai.EmbeddingResponseItem, 0, len(response.Data)),
Model: "hunyuan-embedding",
Usage: model.Usage{TotalTokens: response.EmbeddingUsage.TotalTokens},
}
for _, item := range response.Data {
openAIEmbeddingResponse.Data = append(openAIEmbeddingResponse.Data, openai.EmbeddingResponseItem{
Object: item.Object,
Index: item.Index,
Embedding: item.Embedding,
})
}
return &openAIEmbeddingResponse
}
func responseTencent2OpenAI(response *ChatResponse) *openai.TextResponse {
fullTextResponse := openai.TextResponse{
Id: response.ReqID,
Object: "chat.completion",
Created: helper.GetTimestamp(),
Usage: model.Usage{
@@ -148,7 +210,7 @@ func Handler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, *
return openai.ErrorWrapper(err, "unmarshal_response_body_failed", http.StatusInternalServerError), nil
}
TencentResponse = responseP.Response
if TencentResponse.Error.Code != 0 {
if TencentResponse.Error.Code != "" {
return &model.ErrorWithStatusCode{
Error: model.Error{
Message: TencentResponse.Error.Message,
@@ -195,7 +257,7 @@ func hmacSha256(s, key string) string {
return string(hashed.Sum(nil))
}
func GetSign(req ChatRequest, adaptor *Adaptor, secId, secKey string) string {
func GetSign(req any, adaptor *Adaptor, secId, secKey string) string {
// build canonical request string
host := "hunyuan.tencentcloudapi.com"
httpRequestMethod := "POST"

46
relay/adaptor/tencent/model.go
View File

@@ -35,16 +35,16 @@ type ChatRequest struct {
// 1. 影响输出文本的多样性，取值越大，生成文本的多样性越强。
// 2. 取值区间为 [0.0, 1.0]，未传值时使用各模型推荐值。
// 3. 非必要不建议使用，不合理的取值会影响效果。
TopP *float64 `json:"TopP"`
TopP *float64 `json:"TopP,omitempty"`
// 说明：
// 1. 较高的数值会使输出更加随机，而较低的数值会使其更加集中和确定。
// 2. 取值区间为 [0.0, 2.0]，未传值时使用各模型推荐值。
// 3. 非必要不建议使用，不合理的取值会影响效果。
Temperature *float64 `json:"Temperature"`
Temperature *float64 `json:"Temperature,omitempty"`
}
type Error struct {
Code int `json:"Code"`
Code string `json:"Code"`
Message string `json:"Message"`
}
@@ -61,15 +61,41 @@ type ResponseChoices struct {
}
type ChatResponse struct {
Choices []ResponseChoices `json:"Choices,omitempty"` // 结果
Created int64 `json:"Created,omitempty"` // unix 时间戳的字符串
Id string `json:"Id,omitempty"` // 会话 id
Usage Usage `json:"Usage,omitempty"` // token 数量
Error Error `json:"Error,omitempty"` // 错误信息 注意：此字段可能返回 null，表示取不到有效值
Note string `json:"Note,omitempty"` // 注释
ReqID string `json:"Req_id,omitempty"` // 唯一请求 Id，每次请求都会返回。用于反馈接口入参
Choices []ResponseChoices `json:"Choices,omitempty"` // 结果
Created int64 `json:"Created,omitempty"` // unix 时间戳的字符串
Id string `json:"Id,omitempty"` // 会话 id
Usage Usage `json:"Usage,omitempty"` // token 数量
Error Error `json:"Error,omitempty"` // 错误信息 注意：此字段可能返回 null，表示取不到有效值
Note string `json:"Note,omitempty"` // 注释
ReqID string `json:"RequestId,omitempty"` // 唯一请求 Id，每次请求都会返回。用于反馈接口入参
}
type ChatResponseP struct {
Response ChatResponse `json:"Response,omitempty"`
}
type EmbeddingRequest struct {
InputList []string `json:"InputList"`
}
type EmbeddingData struct {
Embedding []float64 `json:"Embedding"`
Index int `json:"Index"`
Object string `json:"Object"`
}
type EmbeddingUsage struct {
PromptTokens int `json:"PromptTokens"`
TotalTokens int `json:"TotalTokens"`
}
type EmbeddingResponse struct {
Data []EmbeddingData `json:"Data"`
EmbeddingUsage EmbeddingUsage `json:"Usage,omitempty"`
RequestId string `json:"RequestId,omitempty"`
Error Error `json:"Error,omitempty"`
}
type EmbeddingResponseP struct {
Response EmbeddingResponse `json:"Response,omitempty"`
}

9
relay/adaptor/vertexai/gemini/adapter.go
View File

@@ -16,9 +16,12 @@ import (
var ModelList = []string{
"gemini-pro", "gemini-pro-vision",
"gemini-1.5-pro-001", "gemini-1.5-flash-001",
"gemini-1.5-pro-002", "gemini-1.5-flash-002",
"gemini-2.0-flash-exp", "gemini-2.0-flash-thinking-exp",
"gemini-exp-1206",
"gemini-1.5-pro-001", "gemini-1.5-pro-002",
"gemini-1.5-flash-001", "gemini-1.5-flash-002",
"gemini-2.0-flash-exp", "gemini-2.0-flash-001",
"gemini-2.0-flash-lite-preview-02-05",
"gemini-2.0-flash-thinking-exp-01-21",
}
type Adaptor struct {

9
relay/adaptor/xai/constants.go
View File

@@ -1,5 +1,14 @@
package xai
//https://console.x.ai/
var ModelList = []string{
"grok-2",
"grok-vision-beta",
"grok-2-vision-1212",
"grok-2-vision",
"grok-2-vision-latest",
"grok-2-1212",
"grok-2-latest",
"grok-beta",
}

14
relay/adaptor/xunfei/constants.go
View File

@@ -1,12 +1,10 @@
package xunfei
var ModelList = []string{
"SparkDesk",
"SparkDesk-v1.1",
"SparkDesk-v2.1",
"SparkDesk-v3.1",
"SparkDesk-v3.1-128K",
"SparkDesk-v3.5",
"SparkDesk-v3.5-32K",
"SparkDesk-v4.0",
"Spark-Lite",
"Spark-Pro",
"Spark-Pro-128K",
"Spark-Max",
"Spark-Max-32K",
"Spark-4.0-Ultra",
}

97
relay/adaptor/xunfei/domain.go Normal file
View File

@@ -0,0 +1,97 @@
package xunfei
import (
"fmt"
"strings"
)
// https://www.xfyun.cn/doc/spark/Web.html#_1-%E6%8E%A5%E5%8F%A3%E8%AF%B4%E6%98%8E
//Spark4.0 Ultra 请求地址，对应的domain参数为4.0Ultra：
//
//wss://spark-api.xf-yun.com/v4.0/chat
//Spark Max-32K请求地址，对应的domain参数为max-32k
//
//wss://spark-api.xf-yun.com/chat/max-32k
//Spark Max请求地址，对应的domain参数为generalv3.5
//
//wss://spark-api.xf-yun.com/v3.5/chat
//Spark Pro-128K请求地址，对应的domain参数为pro-128k：
//
// wss://spark-api.xf-yun.com/chat/pro-128k
//Spark Pro请求地址，对应的domain参数为generalv3：
//
//wss://spark-api.xf-yun.com/v3.1/chat
//Spark Lite请求地址，对应的domain参数为lite：
//
//wss://spark-api.xf-yun.com/v1.1/chat
// Lite、Pro、Pro-128K、Max、Max-32K和4.0 Ultra
func parseAPIVersionByModelName(modelName string) string {
apiVersion := modelName2APIVersion(modelName)
if apiVersion != "" {
return apiVersion
}
index := strings.IndexAny(modelName, "-")
if index != -1 {
return modelName[index+1:]
}
return ""
}
func modelName2APIVersion(modelName string) string {
switch modelName {
case "Spark-Lite":
return "v1.1"
case "Spark-Pro":
return "v3.1"
case "Spark-Pro-128K":
return "v3.1-128K"
case "Spark-Max":
return "v3.5"
case "Spark-Max-32K":
return "v3.5-32K"
case "Spark-4.0-Ultra":
return "v4.0"
}
return ""
}
// https://www.xfyun.cn/doc/spark/Web.html#_1-%E6%8E%A5%E5%8F%A3%E8%AF%B4%E6%98%8E
func apiVersion2domain(apiVersion string) string {
switch apiVersion {
case "v1.1":
return "lite"
case "v2.1":
return "generalv2"
case "v3.1":
return "generalv3"
case "v3.1-128K":
return "pro-128k"
case "v3.5":
return "generalv3.5"
case "v3.5-32K":
return "max-32k"
case "v4.0":
return "4.0Ultra"
}
return "general" + apiVersion
}
func getXunfeiAuthUrl(apiVersion string, apiKey string, apiSecret string) (string, string) {
var authUrl string
domain := apiVersion2domain(apiVersion)
switch apiVersion {
case "v3.1-128K":
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)
}
return domain, authUrl
}

46
relay/adaptor/xunfei/main.go
View File

@@ -15,6 +15,7 @@ import (
"github.com/gin-gonic/gin"
"github.com/gorilla/websocket"
"github.com/songquanpeng/one-api/common"
"github.com/songquanpeng/one-api/common/helper"
"github.com/songquanpeng/one-api/common/logger"
@@ -270,48 +271,3 @@ func xunfeiMakeRequest(textRequest model.GeneralOpenAIRequest, domain, authUrl,
return dataChan, stopChan, nil
}
func parseAPIVersionByModelName(modelName string) string {
index := strings.IndexAny(modelName, "-")
if index != -1 {
return modelName[index+1:]
}
return ""
}
// https://www.xfyun.cn/doc/spark/Web.html#_1-%E6%8E%A5%E5%8F%A3%E8%AF%B4%E6%98%8E
func apiVersion2domain(apiVersion string) string {
switch apiVersion {
case "v1.1":
return "lite"
case "v2.1":
return "generalv2"
case "v3.1":
return "generalv3"
case "v3.1-128K":
return "pro-128k"
case "v3.5":
return "generalv3.5"
case "v3.5-32K":
return "max-32k"
case "v4.0":
return "4.0Ultra"
}
return "general" + apiVersion
}
func getXunfeiAuthUrl(apiVersion string, apiKey string, apiSecret string) (string, string) {
var authUrl string
domain := apiVersion2domain(apiVersion)
switch apiVersion {
case "v3.1-128K":
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)
}
return domain, authUrl
}

12
relay/adaptor/xunfeiv2/constants.go Normal file
View File

@@ -0,0 +1,12 @@
package xunfeiv2
// https://www.xfyun.cn/doc/spark/HTTP%E8%B0%83%E7%94%A8%E6%96%87%E6%A1%A3.html#_3-%E8%AF%B7%E6%B1%82%E8%AF%B4%E6%98%8E
var ModelList = []string{
"lite",
"generalv3",
"pro-128k",
"generalv3.5",
"max-32k",
"4.0Ultra",
}

13
relay/adaptor/zhipu/constants.go
View File

@@ -1,7 +1,14 @@
package zhipu
// https://open.bigmodel.cn/pricing
var ModelList = []string{
"chatglm_turbo", "chatglm_pro", "chatglm_std", "chatglm_lite",
"glm-4", "glm-4v", "glm-3-turbo", "embedding-2",
"cogview-3",
"glm-zero-preview", "glm-4-plus", "glm-4-0520", "glm-4-airx",
"glm-4-air", "glm-4-long", "glm-4-flashx", "glm-4-flash",
"glm-4", "glm-3-turbo",
"glm-4v-plus", "glm-4v", "glm-4v-flash",
"cogview-3-plus", "cogview-3", "cogview-3-flash",
"cogviewx", "cogviewx-flash",
"charglm-4", "emohaa", "codegeex-4",
"embedding-2", "embedding-3",
}

14
relay/billing/billing.go
View File

@@ -3,6 +3,7 @@ package billing
import (
"context"
"fmt"
"github.com/songquanpeng/one-api/common/logger"
"github.com/songquanpeng/one-api/model"
)
@@ -31,8 +32,17 @@ func PostConsumeQuota(ctx context.Context, tokenId int, quotaDelta int64, totalQ
}
// totalQuota is total quota consumed
if totalQuota != 0 {
logContent := fmt.Sprintf("模型倍率 %.2f，分组倍率 %.2f", modelRatio, groupRatio)
model.RecordConsumeLog(ctx, userId, channelId, int(totalQuota), 0, modelName, tokenName, totalQuota, logContent)
logContent := fmt.Sprintf("倍率：%.2f × %.2f", modelRatio, groupRatio)
model.RecordConsumeLog(ctx, &model.Log{
UserId: userId,
ChannelId: channelId,
PromptTokens: int(totalQuota),
CompletionTokens: 0,
ModelName: modelName,
TokenName: tokenName,
Quota: int(totalQuota),
Content: logContent,
})
model.UpdateUserUsedQuotaAndRequestCount(userId, totalQuota)
model.UpdateChannelUsedQuota(channelId, totalQuota)
}

6
relay/billing/ratio/group.go
View File

@@ -3,8 +3,10 @@ package ratio
import (
"encoding/json"
"github.com/songquanpeng/one-api/common/logger"
"sync"
)
var groupRatioLock sync.RWMutex
var GroupRatio = map[string]float64{
"default": 1,
"vip": 1,
@@ -20,11 +22,15 @@ func GroupRatio2JSONString() string {
}
func UpdateGroupRatioByJSONString(jsonStr string) error {
groupRatioLock.Lock()
defer groupRatioLock.Unlock()
GroupRatio = make(map[string]float64)
return json.Unmarshal([]byte(jsonStr), &GroupRatio)
}
func GetGroupRatio(name string) float64 {
groupRatioLock.RLock()
defer groupRatioLock.RUnlock()
ratio, ok := GroupRatio[name]
if !ok {
logger.SysError("group ratio not found: " + name)

521
relay/billing/ratio/model.go
View File

@@ -4,16 +4,20 @@ import (
"encoding/json"
"fmt"
"strings"
"sync"
"github.com/songquanpeng/one-api/common/logger"
)
const (
USD2RMB = 7
USD = 500 // $0.002 = 1 -> $1 = 500
RMB = USD / USD2RMB
USD2RMB = 7
USD = 500 // $0.002 = 1 -> $1 = 500
MILLI_USD = 1.0 / 1000 * USD
RMB = USD / USD2RMB
)
var modelRatioLock sync.RWMutex
// ModelRatio
// https://platform.openai.com/docs/models/model-endpoint-compatibility
// https://cloud.baidu.com/doc/WENXINWORKSHOP/s/Blfmc9dlf
@@ -55,6 +59,8 @@ var ModelRatio = map[string]float64{
"o1-preview-2024-09-12": 7.5,
"o1-mini": 1.5, // $3.00 / 1M input tokens
"o1-mini-2024-09-12": 1.5,
"o3-mini": 1.5, // $3.00 / 1M input tokens
"o3-mini-2025-01-31": 1.5,
"davinci-002": 1, // $0.002 / 1K tokens
"babbage-002": 0.2, // $0.0004 / 1K tokens
"text-ada-001": 0.2,
@@ -81,15 +87,17 @@ var ModelRatio = map[string]float64{
"text-moderation-latest": 0.1,
"dall-e-2": 0.02 * USD, // $0.016 - $0.020 / image
"dall-e-3": 0.04 * USD, // $0.040 - $0.120 / image
// https://www.anthropic.com/api#pricing
// https://docs.anthropic.com/en/docs/about-claude/models
"claude-instant-1.2": 0.8 / 1000 * USD,
"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-3-5-haiku-latest": 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-3-5-sonnet-latest": 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,
@@ -109,114 +117,162 @@ var ModelRatio = map[string]float64{
"bge-large-en": 0.002 * RMB,
"tao-8k": 0.002 * RMB,
// https://ai.google.dev/pricing
"gemini-pro": 1, // $0.00025 / 1k characters -> $0.001 / 1k tokens
"gemini-1.0-pro": 1,
"gemini-1.5-pro": 1,
"gemini-1.5-pro-001": 1,
"gemini-1.5-flash": 1,
"gemini-1.5-flash-001": 1,
"gemini-2.0-flash-exp": 1,
"gemini-2.0-flash-thinking-exp": 1,
"aqa": 1,
// https://cloud.google.com/vertex-ai/generative-ai/pricing
// "gemma-2-2b-it": 0,
// "gemma-2-9b-it": 0,
// "gemma-2-27b-it": 0,
"gemini-pro": 0.25 * MILLI_USD, // $0.00025 / 1k characters -> $0.001 / 1k tokens
"gemini-1.0-pro": 0.125 * MILLI_USD,
"gemini-1.5-pro": 1.25 * MILLI_USD,
"gemini-1.5-pro-001": 1.25 * MILLI_USD,
"gemini-1.5-pro-experimental": 1.25 * MILLI_USD,
"gemini-1.5-flash": 0.075 * MILLI_USD,
"gemini-1.5-flash-001": 0.075 * MILLI_USD,
"gemini-1.5-flash-8b": 0.0375 * MILLI_USD,
"gemini-2.0-flash-exp": 0.075 * MILLI_USD,
"gemini-2.0-flash": 0.15 * MILLI_USD,
"gemini-2.0-flash-001": 0.15 * MILLI_USD,
"gemini-2.0-flash-lite-preview-02-05": 0.075 * MILLI_USD,
"gemini-2.0-flash-thinking-exp-01-21": 0.075 * MILLI_USD,
"gemini-2.0-pro-exp-02-05": 1.25 * MILLI_USD,
"aqa": 1,
// https://open.bigmodel.cn/pricing
"glm-4": 0.1 * RMB,
"glm-4v": 0.1 * RMB,
"glm-3-turbo": 0.005 * RMB,
"embedding-2": 0.0005 * RMB,
"chatglm_turbo": 0.3572, // ￥0.005 / 1k tokens
"chatglm_pro": 0.7143, // ￥0.01 / 1k tokens
"chatglm_std": 0.3572, // ￥0.005 / 1k tokens
"chatglm_lite": 0.1429, // ￥0.002 / 1k tokens
"cogview-3": 0.25 * RMB,
"glm-zero-preview": 0.01 * RMB,
"glm-4-plus": 0.05 * RMB,
"glm-4-0520": 0.1 * RMB,
"glm-4-airx": 0.01 * RMB,
"glm-4-air": 0.0005 * RMB,
"glm-4-long": 0.001 * RMB,
"glm-4-flashx": 0.0001 * RMB,
"glm-4-flash": 0,
"glm-4": 0.1 * RMB, // deprecated model, available until 2025/06
"glm-3-turbo": 0.001 * RMB, // deprecated model, available until 2025/06
"glm-4v-plus": 0.004 * RMB,
"glm-4v": 0.05 * RMB,
"glm-4v-flash": 0,
"cogview-3-plus": 0.06 * RMB,
"cogview-3": 0.1 * RMB,
"cogview-3-flash": 0,
"cogviewx": 0.5 * RMB,
"cogviewx-flash": 0,
"charglm-4": 0.001 * RMB,
"emohaa": 0.015 * RMB,
"codegeex-4": 0.0001 * RMB,
"embedding-2": 0.0005 * RMB,
"embedding-3": 0.0005 * RMB,
// https://help.aliyun.com/zh/dashscope/developer-reference/tongyi-thousand-questions-metering-and-billing
"qwen-turbo": 1.4286, // ￥0.02 / 1k tokens
"qwen-turbo-latest": 1.4286,
"qwen-plus": 1.4286,
"qwen-plus-latest": 1.4286,
"qwen-max": 1.4286,
"qwen-max-latest": 1.4286,
"qwen-max-longcontext": 1.4286,
"qwen-vl-max": 1.4286,
"qwen-vl-max-latest": 1.4286,
"qwen-vl-plus": 1.4286,
"qwen-vl-plus-latest": 1.4286,
"qwen-vl-ocr": 1.4286,
"qwen-vl-ocr-latest": 1.4286,
"qwen-audio-turbo": 1.4286,
"qwen-math-plus": 1.4286,
"qwen-math-plus-latest": 1.4286,
"qwen-math-turbo": 1.4286,
"qwen-math-turbo-latest": 1.4286,
"qwen-coder-plus": 1.4286,
"qwen-coder-plus-latest": 1.4286,
"qwen-coder-turbo": 1.4286,
"qwen-coder-turbo-latest": 1.4286,
"qwq-32b-preview": 1.4286,
"qwen2.5-72b-instruct": 1.4286,
"qwen2.5-32b-instruct": 1.4286,
"qwen2.5-14b-instruct": 1.4286,
"qwen2.5-7b-instruct": 1.4286,
"qwen2.5-3b-instruct": 1.4286,
"qwen2.5-1.5b-instruct": 1.4286,
"qwen2.5-0.5b-instruct": 1.4286,
"qwen2-72b-instruct": 1.4286,
"qwen2-57b-a14b-instruct": 1.4286,
"qwen2-7b-instruct": 1.4286,
"qwen2-1.5b-instruct": 1.4286,
"qwen2-0.5b-instruct": 1.4286,
"qwen1.5-110b-chat": 1.4286,
"qwen1.5-72b-chat": 1.4286,
"qwen1.5-32b-chat": 1.4286,
"qwen1.5-14b-chat": 1.4286,
"qwen1.5-7b-chat": 1.4286,
"qwen1.5-1.8b-chat": 1.4286,
"qwen1.5-0.5b-chat": 1.4286,
"qwen-72b-chat": 1.4286,
"qwen-14b-chat": 1.4286,
"qwen-7b-chat": 1.4286,
"qwen-1.8b-chat": 1.4286,
"qwen-1.8b-longcontext-chat": 1.4286,
"qwen2-vl-7b-instruct": 1.4286,
"qwen2-vl-2b-instruct": 1.4286,
"qwen-vl-v1": 1.4286,
"qwen-vl-chat-v1": 1.4286,
"qwen2-audio-instruct": 1.4286,
"qwen-audio-chat": 1.4286,
"qwen2.5-math-72b-instruct": 1.4286,
"qwen2.5-math-7b-instruct": 1.4286,
"qwen2.5-math-1.5b-instruct": 1.4286,
"qwen2-math-72b-instruct": 1.4286,
"qwen2-math-7b-instruct": 1.4286,
"qwen2-math-1.5b-instruct": 1.4286,
"qwen2.5-coder-32b-instruct": 1.4286,
"qwen2.5-coder-14b-instruct": 1.4286,
"qwen2.5-coder-7b-instruct": 1.4286,
"qwen2.5-coder-3b-instruct": 1.4286,
"qwen2.5-coder-1.5b-instruct": 1.4286,
"qwen2.5-coder-0.5b-instruct": 1.4286,
"text-embedding-v1": 0.05, // ￥0.0007 / 1k tokens
"text-embedding-v3": 0.05,
"text-embedding-v2": 0.05,
"text-embedding-async-v2": 0.05,
"text-embedding-async-v1": 0.05,
"ali-stable-diffusion-xl": 8.00,
"ali-stable-diffusion-v1.5": 8.00,
"wanx-v1": 8.00,
"SparkDesk": 1.2858, // ￥0.018 / 1k tokens
"SparkDesk-v1.1": 1.2858, // ￥0.018 / 1k tokens
"SparkDesk-v2.1": 1.2858, // ￥0.018 / 1k tokens
"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
"embedding_s1_v1": 0.0715, // ¥0.001 / 1k tokens
"semantic_similarity_s1_v1": 0.0715, // ¥0.001 / 1k tokens
"hunyuan": 7.143, // ¥0.1 / 1k tokens // https://cloud.tencent.com/document/product/1729/97731#e0e6be58-60c8-469f-bdeb-6c264ce3b4d0
"ChatStd": 0.01 * RMB,
"ChatPro": 0.1 * RMB,
"qwen-turbo": 0.0003 * RMB,
"qwen-turbo-latest": 0.0003 * RMB,
"qwen-plus": 0.0008 * RMB,
"qwen-plus-latest": 0.0008 * RMB,
"qwen-max": 0.0024 * RMB,
"qwen-max-latest": 0.0024 * RMB,
"qwen-max-longcontext": 0.0005 * RMB,
"qwen-vl-max": 0.003 * RMB,
"qwen-vl-max-latest": 0.003 * RMB,
"qwen-vl-plus": 0.0015 * RMB,
"qwen-vl-plus-latest": 0.0015 * RMB,
"qwen-vl-ocr": 0.005 * RMB,
"qwen-vl-ocr-latest": 0.005 * RMB,
"qwen-audio-turbo": 1.4286,
"qwen-math-plus": 0.004 * RMB,
"qwen-math-plus-latest": 0.004 * RMB,
"qwen-math-turbo": 0.002 * RMB,
"qwen-math-turbo-latest": 0.002 * RMB,
"qwen-coder-plus": 0.0035 * RMB,
"qwen-coder-plus-latest": 0.0035 * RMB,
"qwen-coder-turbo": 0.002 * RMB,
"qwen-coder-turbo-latest": 0.002 * RMB,
"qwen-mt-plus": 0.015 * RMB,
"qwen-mt-turbo": 0.001 * RMB,
"qwq-32b-preview": 0.002 * RMB,
"qwen2.5-72b-instruct": 0.004 * RMB,
"qwen2.5-32b-instruct": 0.03 * RMB,
"qwen2.5-14b-instruct": 0.001 * RMB,
"qwen2.5-7b-instruct": 0.0005 * RMB,
"qwen2.5-3b-instruct": 0.006 * RMB,
"qwen2.5-1.5b-instruct": 0.0003 * RMB,
"qwen2.5-0.5b-instruct": 0.0003 * RMB,
"qwen2-72b-instruct": 0.004 * RMB,
"qwen2-57b-a14b-instruct": 0.0035 * RMB,
"qwen2-7b-instruct": 0.001 * RMB,
"qwen2-1.5b-instruct": 0.001 * RMB,
"qwen2-0.5b-instruct": 0.001 * RMB,
"qwen1.5-110b-chat": 0.007 * RMB,
"qwen1.5-72b-chat": 0.005 * RMB,
"qwen1.5-32b-chat": 0.0035 * RMB,
"qwen1.5-14b-chat": 0.002 * RMB,
"qwen1.5-7b-chat": 0.001 * RMB,
"qwen1.5-1.8b-chat": 0.001 * RMB,
"qwen1.5-0.5b-chat": 0.001 * RMB,
"qwen-72b-chat": 0.02 * RMB,
"qwen-14b-chat": 0.008 * RMB,
"qwen-7b-chat": 0.006 * RMB,
"qwen-1.8b-chat": 0.006 * RMB,
"qwen-1.8b-longcontext-chat": 0.006 * RMB,
"qvq-72b-preview": 0.012 * RMB,
"qwen2.5-vl-72b-instruct": 0.016 * RMB,
"qwen2.5-vl-7b-instruct": 0.002 * RMB,
"qwen2.5-vl-3b-instruct": 0.0012 * RMB,
"qwen2-vl-7b-instruct": 0.016 * RMB,
"qwen2-vl-2b-instruct": 0.002 * RMB,
"qwen-vl-v1": 0.002 * RMB,
"qwen-vl-chat-v1": 0.002 * RMB,
"qwen2-audio-instruct": 0.002 * RMB,
"qwen-audio-chat": 0.002 * RMB,
"qwen2.5-math-72b-instruct": 0.004 * RMB,
"qwen2.5-math-7b-instruct": 0.001 * RMB,
"qwen2.5-math-1.5b-instruct": 0.001 * RMB,
"qwen2-math-72b-instruct": 0.004 * RMB,
"qwen2-math-7b-instruct": 0.001 * RMB,
"qwen2-math-1.5b-instruct": 0.001 * RMB,
"qwen2.5-coder-32b-instruct": 0.002 * RMB,
"qwen2.5-coder-14b-instruct": 0.002 * RMB,
"qwen2.5-coder-7b-instruct": 0.001 * RMB,
"qwen2.5-coder-3b-instruct": 0.001 * RMB,
"qwen2.5-coder-1.5b-instruct": 0.001 * RMB,
"qwen2.5-coder-0.5b-instruct": 0.001 * RMB,
"text-embedding-v1": 0.0007 * RMB, // ￥0.0007 / 1k tokens
"text-embedding-v3": 0.0007 * RMB,
"text-embedding-v2": 0.0007 * RMB,
"text-embedding-async-v2": 0.0007 * RMB,
"text-embedding-async-v1": 0.0007 * RMB,
"ali-stable-diffusion-xl": 8.00,
"ali-stable-diffusion-v1.5": 8.00,
"wanx-v1": 8.00,
"deepseek-r1": 0.002 * RMB,
"deepseek-v3": 0.001 * RMB,
"deepseek-r1-distill-qwen-1.5b": 0.001 * RMB,
"deepseek-r1-distill-qwen-7b": 0.0005 * RMB,
"deepseek-r1-distill-qwen-14b": 0.001 * RMB,
"deepseek-r1-distill-qwen-32b": 0.002 * RMB,
"deepseek-r1-distill-llama-8b": 0.0005 * RMB,
"deepseek-r1-distill-llama-70b": 0.004 * RMB,
"SparkDesk": 1.2858, // ￥0.018 / 1k tokens
"SparkDesk-v1.1": 1.2858, // ￥0.018 / 1k tokens
"SparkDesk-v2.1": 1.2858, // ￥0.018 / 1k tokens
"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
"embedding_s1_v1": 0.0715, // ¥0.001 / 1k tokens
"semantic_similarity_s1_v1": 0.0715, // ¥0.001 / 1k tokens
// https://cloud.tencent.com/document/product/1729/97731#e0e6be58-60c8-469f-bdeb-6c264ce3b4d0
"hunyuan-turbo": 0.015 * RMB,
"hunyuan-large": 0.004 * RMB,
"hunyuan-large-longcontext": 0.006 * RMB,
"hunyuan-standard": 0.0008 * RMB,
"hunyuan-standard-256K": 0.0005 * RMB,
"hunyuan-translation-lite": 0.005 * RMB,
"hunyuan-role": 0.004 * RMB,
"hunyuan-functioncall": 0.004 * RMB,
"hunyuan-code": 0.004 * RMB,
"hunyuan-turbo-vision": 0.08 * RMB,
"hunyuan-vision": 0.018 * RMB,
"hunyuan-embedding": 0.0007 * RMB,
// https://platform.moonshot.cn/pricing
"moonshot-v1-8k": 0.012 * RMB,
"moonshot-v1-32k": 0.024 * RMB,
@@ -279,8 +335,8 @@ var ModelRatio = map[string]float64{
"command-r": 0.5 / 1000 * USD,
"command-r-plus": 3.0 / 1000 * USD,
// https://platform.deepseek.com/api-docs/pricing/
"deepseek-chat": 1.0 / 1000 * RMB,
"deepseek-coder": 1.0 / 1000 * RMB,
"deepseek-chat": 0.14 * MILLI_USD,
"deepseek-reasoner": 0.55 * MILLI_USD,
// https://www.deepl.com/pro?cta=header-prices
"deepl-zh": 25.0 / 1000 * USD,
"deepl-en": 25.0 / 1000 * USD,
@@ -331,12 +387,249 @@ var ModelRatio = map[string]float64{
"mistralai/mistral-7b-instruct-v0.2": 0.050 * USD,
"mistralai/mistral-7b-v0.1": 0.050 * USD,
"mistralai/mixtral-8x7b-instruct-v0.1": 0.300 * USD,
//https://openrouter.ai/models
"01-ai/yi-large": 1.5,
"aetherwiing/mn-starcannon-12b": 0.6,
"ai21/jamba-1-5-large": 4.0,
"ai21/jamba-1-5-mini": 0.2,
"ai21/jamba-instruct": 0.35,
"aion-labs/aion-1.0": 6.0,
"aion-labs/aion-1.0-mini": 1.2,
"aion-labs/aion-rp-llama-3.1-8b": 0.1,
"allenai/llama-3.1-tulu-3-405b": 5.0,
"alpindale/goliath-120b": 4.6875,
"alpindale/magnum-72b": 1.125,
"amazon/nova-lite-v1": 0.12,
"amazon/nova-micro-v1": 0.07,
"amazon/nova-pro-v1": 1.6,
"anthracite-org/magnum-v2-72b": 1.5,
"anthracite-org/magnum-v4-72b": 1.125,
"anthropic/claude-2": 12.0,
"anthropic/claude-2.0": 12.0,
"anthropic/claude-2.0:beta": 12.0,
"anthropic/claude-2.1": 12.0,
"anthropic/claude-2.1:beta": 12.0,
"anthropic/claude-2:beta": 12.0,
"anthropic/claude-3-haiku": 0.625,
"anthropic/claude-3-haiku:beta": 0.625,
"anthropic/claude-3-opus": 37.5,
"anthropic/claude-3-opus:beta": 37.5,
"anthropic/claude-3-sonnet": 7.5,
"anthropic/claude-3-sonnet:beta": 7.5,
"anthropic/claude-3.5-haiku": 2.0,
"anthropic/claude-3.5-haiku-20241022": 2.0,
"anthropic/claude-3.5-haiku-20241022:beta": 2.0,
"anthropic/claude-3.5-haiku:beta": 2.0,
"anthropic/claude-3.5-sonnet": 7.5,
"anthropic/claude-3.5-sonnet-20240620": 7.5,
"anthropic/claude-3.5-sonnet-20240620:beta": 7.5,
"anthropic/claude-3.5-sonnet:beta": 7.5,
"cognitivecomputations/dolphin-mixtral-8x22b": 0.45,
"cognitivecomputations/dolphin-mixtral-8x7b": 0.25,
"cohere/command": 0.95,
"cohere/command-r": 0.7125,
"cohere/command-r-03-2024": 0.7125,
"cohere/command-r-08-2024": 0.285,
"cohere/command-r-plus": 7.125,
"cohere/command-r-plus-04-2024": 7.125,
"cohere/command-r-plus-08-2024": 4.75,
"cohere/command-r7b-12-2024": 0.075,
"databricks/dbrx-instruct": 0.6,
"deepseek/deepseek-chat": 0.445,
"deepseek/deepseek-chat-v2.5": 1.0,
"deepseek/deepseek-chat:free": 0.0,
"deepseek/deepseek-r1": 1.2,
"deepseek/deepseek-r1-distill-llama-70b": 0.345,
"deepseek/deepseek-r1-distill-llama-70b:free": 0.0,
"deepseek/deepseek-r1-distill-llama-8b": 0.02,
"deepseek/deepseek-r1-distill-qwen-1.5b": 0.09,
"deepseek/deepseek-r1-distill-qwen-14b": 0.075,
"deepseek/deepseek-r1-distill-qwen-32b": 0.09,
"deepseek/deepseek-r1:free": 0.0,
"eva-unit-01/eva-llama-3.33-70b": 3.0,
"eva-unit-01/eva-qwen-2.5-32b": 1.7,
"eva-unit-01/eva-qwen-2.5-72b": 3.0,
"google/gemini-2.0-flash-001": 0.2,
"google/gemini-2.0-flash-exp:free": 0.0,
"google/gemini-2.0-flash-lite-preview-02-05:free": 0.0,
"google/gemini-2.0-flash-thinking-exp-1219:free": 0.0,
"google/gemini-2.0-flash-thinking-exp:free": 0.0,
"google/gemini-2.0-pro-exp-02-05:free": 0.0,
"google/gemini-exp-1206:free": 0.0,
"google/gemini-flash-1.5": 0.15,
"google/gemini-flash-1.5-8b": 0.075,
"google/gemini-flash-1.5-8b-exp": 0.0,
"google/gemini-pro": 0.75,
"google/gemini-pro-1.5": 2.5,
"google/gemini-pro-vision": 0.75,
"google/gemma-2-27b-it": 0.135,
"google/gemma-2-9b-it": 0.03,
"google/gemma-2-9b-it:free": 0.0,
"google/gemma-7b-it": 0.075,
"google/learnlm-1.5-pro-experimental:free": 0.0,
"google/palm-2-chat-bison": 1.0,
"google/palm-2-chat-bison-32k": 1.0,
"google/palm-2-codechat-bison": 1.0,
"google/palm-2-codechat-bison-32k": 1.0,
"gryphe/mythomax-l2-13b": 0.0325,
"gryphe/mythomax-l2-13b:free": 0.0,
"huggingfaceh4/zephyr-7b-beta:free": 0.0,
"infermatic/mn-inferor-12b": 0.6,
"inflection/inflection-3-pi": 5.0,
"inflection/inflection-3-productivity": 5.0,
"jondurbin/airoboros-l2-70b": 0.25,
"liquid/lfm-3b": 0.01,
"liquid/lfm-40b": 0.075,
"liquid/lfm-7b": 0.005,
"mancer/weaver": 1.125,
"meta-llama/llama-2-13b-chat": 0.11,
"meta-llama/llama-2-70b-chat": 0.45,
"meta-llama/llama-3-70b-instruct": 0.2,
"meta-llama/llama-3-8b-instruct": 0.03,
"meta-llama/llama-3-8b-instruct:free": 0.0,
"meta-llama/llama-3.1-405b": 1.0,
"meta-llama/llama-3.1-405b-instruct": 0.4,
"meta-llama/llama-3.1-70b-instruct": 0.15,
"meta-llama/llama-3.1-8b-instruct": 0.025,
"meta-llama/llama-3.2-11b-vision-instruct": 0.0275,
"meta-llama/llama-3.2-11b-vision-instruct:free": 0.0,
"meta-llama/llama-3.2-1b-instruct": 0.005,
"meta-llama/llama-3.2-3b-instruct": 0.0125,
"meta-llama/llama-3.2-90b-vision-instruct": 0.8,
"meta-llama/llama-3.3-70b-instruct": 0.15,
"meta-llama/llama-3.3-70b-instruct:free": 0.0,
"meta-llama/llama-guard-2-8b": 0.1,
"microsoft/phi-3-medium-128k-instruct": 0.5,
"microsoft/phi-3-medium-128k-instruct:free": 0.0,
"microsoft/phi-3-mini-128k-instruct": 0.05,
"microsoft/phi-3-mini-128k-instruct:free": 0.0,
"microsoft/phi-3.5-mini-128k-instruct": 0.05,
"microsoft/phi-4": 0.07,
"microsoft/wizardlm-2-7b": 0.035,
"microsoft/wizardlm-2-8x22b": 0.25,
"minimax/minimax-01": 0.55,
"mistralai/codestral-2501": 0.45,
"mistralai/codestral-mamba": 0.125,
"mistralai/ministral-3b": 0.02,
"mistralai/ministral-8b": 0.05,
"mistralai/mistral-7b-instruct": 0.0275,
"mistralai/mistral-7b-instruct-v0.1": 0.1,
"mistralai/mistral-7b-instruct-v0.3": 0.0275,
"mistralai/mistral-7b-instruct:free": 0.0,
"mistralai/mistral-large": 3.0,
"mistralai/mistral-large-2407": 3.0,
"mistralai/mistral-large-2411": 3.0,
"mistralai/mistral-medium": 4.05,
"mistralai/mistral-nemo": 0.04,
"mistralai/mistral-nemo:free": 0.0,
"mistralai/mistral-small": 0.3,
"mistralai/mistral-small-24b-instruct-2501": 0.07,
"mistralai/mistral-small-24b-instruct-2501:free": 0.0,
"mistralai/mistral-tiny": 0.125,
"mistralai/mixtral-8x22b-instruct": 0.45,
"mistralai/mixtral-8x7b": 0.3,
"mistralai/mixtral-8x7b-instruct": 0.12,
"mistralai/pixtral-12b": 0.05,
"mistralai/pixtral-large-2411": 3.0,
"neversleep/llama-3-lumimaid-70b": 2.25,
"neversleep/llama-3-lumimaid-8b": 0.5625,
"neversleep/llama-3-lumimaid-8b:extended": 0.5625,
"neversleep/llama-3.1-lumimaid-70b": 2.25,
"neversleep/llama-3.1-lumimaid-8b": 0.5625,
"neversleep/noromaid-20b": 1.125,
"nothingiisreal/mn-celeste-12b": 0.6,
"nousresearch/hermes-2-pro-llama-3-8b": 0.02,
"nousresearch/hermes-3-llama-3.1-405b": 0.4,
"nousresearch/hermes-3-llama-3.1-70b": 0.15,
"nousresearch/nous-hermes-2-mixtral-8x7b-dpo": 0.3,
"nousresearch/nous-hermes-llama2-13b": 0.085,
"nvidia/llama-3.1-nemotron-70b-instruct": 0.15,
"nvidia/llama-3.1-nemotron-70b-instruct:free": 0.0,
"openai/chatgpt-4o-latest": 7.5,
"openai/gpt-3.5-turbo": 0.75,
"openai/gpt-3.5-turbo-0125": 0.75,
"openai/gpt-3.5-turbo-0613": 1.0,
"openai/gpt-3.5-turbo-1106": 1.0,
"openai/gpt-3.5-turbo-16k": 2.0,
"openai/gpt-3.5-turbo-instruct": 1.0,
"openai/gpt-4": 30.0,
"openai/gpt-4-0314": 30.0,
"openai/gpt-4-1106-preview": 15.0,
"openai/gpt-4-32k": 60.0,
"openai/gpt-4-32k-0314": 60.0,
"openai/gpt-4-turbo": 15.0,
"openai/gpt-4-turbo-preview": 15.0,
"openai/gpt-4o": 5.0,
"openai/gpt-4o-2024-05-13": 7.5,
"openai/gpt-4o-2024-08-06": 5.0,
"openai/gpt-4o-2024-11-20": 5.0,
"openai/gpt-4o-mini": 0.3,
"openai/gpt-4o-mini-2024-07-18": 0.3,
"openai/gpt-4o:extended": 9.0,
"openai/o1": 30.0,
"openai/o1-mini": 2.2,
"openai/o1-mini-2024-09-12": 2.2,
"openai/o1-preview": 30.0,
"openai/o1-preview-2024-09-12": 30.0,
"openai/o3-mini": 2.2,
"openai/o3-mini-high": 2.2,
"openchat/openchat-7b": 0.0275,
"openchat/openchat-7b:free": 0.0,
"openrouter/auto": -500000.0,
"perplexity/llama-3.1-sonar-huge-128k-online": 2.5,
"perplexity/llama-3.1-sonar-large-128k-chat": 0.5,
"perplexity/llama-3.1-sonar-large-128k-online": 0.5,
"perplexity/llama-3.1-sonar-small-128k-chat": 0.1,
"perplexity/llama-3.1-sonar-small-128k-online": 0.1,
"perplexity/sonar": 0.5,
"perplexity/sonar-reasoning": 2.5,
"pygmalionai/mythalion-13b": 0.6,
"qwen/qvq-72b-preview": 0.25,
"qwen/qwen-2-72b-instruct": 0.45,
"qwen/qwen-2-7b-instruct": 0.027,
"qwen/qwen-2-7b-instruct:free": 0.0,
"qwen/qwen-2-vl-72b-instruct": 0.2,
"qwen/qwen-2-vl-7b-instruct": 0.05,
"qwen/qwen-2.5-72b-instruct": 0.2,
"qwen/qwen-2.5-7b-instruct": 0.025,
"qwen/qwen-2.5-coder-32b-instruct": 0.08,
"qwen/qwen-max": 3.2,
"qwen/qwen-plus": 0.6,
"qwen/qwen-turbo": 0.1,
"qwen/qwen-vl-plus:free": 0.0,
"qwen/qwen2.5-vl-72b-instruct:free": 0.0,
"qwen/qwq-32b-preview": 0.09,
"raifle/sorcererlm-8x22b": 2.25,
"sao10k/fimbulvetr-11b-v2": 0.6,
"sao10k/l3-euryale-70b": 0.4,
"sao10k/l3-lunaris-8b": 0.03,
"sao10k/l3.1-70b-hanami-x1": 1.5,
"sao10k/l3.1-euryale-70b": 0.4,
"sao10k/l3.3-euryale-70b": 0.4,
"sophosympatheia/midnight-rose-70b": 0.4,
"sophosympatheia/rogue-rose-103b-v0.2:free": 0.0,
"teknium/openhermes-2.5-mistral-7b": 0.085,
"thedrummer/rocinante-12b": 0.25,
"thedrummer/unslopnemo-12b": 0.25,
"undi95/remm-slerp-l2-13b": 0.6,
"undi95/toppy-m-7b": 0.035,
"undi95/toppy-m-7b:free": 0.0,
"x-ai/grok-2-1212": 5.0,
"x-ai/grok-2-vision-1212": 5.0,
"x-ai/grok-beta": 7.5,
"x-ai/grok-vision-beta": 7.5,
"xwin-lm/xwin-lm-70b": 1.875,
}
var CompletionRatio = map[string]float64{
// aws llama3
"llama3-8b-8192(33)": 0.0006 / 0.0003,
"llama3-70b-8192(33)": 0.0035 / 0.00265,
// whisper
"whisper-1": 0, // only count input tokens
// deepseek
"deepseek-chat": 0.28 / 0.14,
"deepseek-reasoner": 2.19 / 0.55,
}
var (
@@ -384,11 +677,15 @@ func ModelRatio2JSONString() string {
}
func UpdateModelRatioByJSONString(jsonStr string) error {
modelRatioLock.Lock()
defer modelRatioLock.Unlock()
ModelRatio = make(map[string]float64)
return json.Unmarshal([]byte(jsonStr), &ModelRatio)
}
func GetModelRatio(name string, channelType int) float64 {
modelRatioLock.RLock()
defer modelRatioLock.RUnlock()
if strings.HasPrefix(name, "qwen-") && strings.HasSuffix(name, "-internet") {
name = strings.TrimSuffix(name, "-internet")
}

5
relay/channeltype/define.go
View File

@@ -48,6 +48,11 @@ const (
SiliconFlow
XAI
Replicate
BaiduV2
XunfeiV2
AliBailian
OpenAICompatible
GeminiOpenAICompatible
CozeV3
Dummy
)

8
relay/channeltype/url.go
View File

@@ -48,7 +48,13 @@ var ChannelBaseURLs = []string{
"https://api.siliconflow.cn", // 44
"https://api.x.ai", // 45
"https://api.replicate.com/v1/models/", // 46
"https://api.coze.cn", // 47
"https://qianfan.baidubce.com", // 47
"https://spark-api-open.xf-yun.com", // 48
"https://dashscope.aliyuncs.com", // 49
"", // 50
"https://generativelanguage.googleapis.com/v1beta/openai/", // 51
"https://api.coze.cn", // 52
}
func init() {

25
relay/controller/helper.go
View File

@@ -4,12 +4,15 @@ import (
"context"
"errors"
"fmt"
"github.com/songquanpeng/one-api/relay/constant/role"
"math"
"net/http"
"strings"
"github.com/songquanpeng/one-api/common/helper"
"github.com/songquanpeng/one-api/relay/constant/role"
"github.com/gin-gonic/gin"
"github.com/songquanpeng/one-api/common"
"github.com/songquanpeng/one-api/common/config"
"github.com/songquanpeng/one-api/common/logger"
@@ -119,12 +122,20 @@ func postConsumeQuota(ctx context.Context, usage *relaymodel.Usage, meta *meta.M
if err != nil {
logger.Error(ctx, "error update user quota cache: "+err.Error())
}
var extraLog string
if systemPromptReset {
extraLog = " （注意系统提示词已被重置）"
}
logContent := fmt.Sprintf("模型倍率 %.2f，分组倍率 %.2f，补全倍率 %.2f%s", modelRatio, groupRatio, completionRatio, extraLog)
model.RecordConsumeLog(ctx, meta.UserId, meta.ChannelId, promptTokens, completionTokens, textRequest.Model, meta.TokenName, quota, logContent)
logContent := fmt.Sprintf("倍率：%.2f × %.2f × %.2f", modelRatio, groupRatio, completionRatio)
model.RecordConsumeLog(ctx, &model.Log{
UserId: meta.UserId,
ChannelId: meta.ChannelId,
PromptTokens: promptTokens,
CompletionTokens: completionTokens,
ModelName: textRequest.Model,
TokenName: meta.TokenName,
Quota: int(quota),
Content: logContent,
IsStream: meta.IsStream,
ElapsedTime: helper.CalcElapsedTime(meta.StartTime),
SystemPromptReset: systemPromptReset,
})
model.UpdateUserUsedQuotaAndRequestCount(meta.UserId, quota)
model.UpdateChannelUsedQuota(meta.ChannelId, quota)
}

14
relay/controller/image.go
View File

@@ -10,6 +10,7 @@ import (
"net/http"
"github.com/gin-gonic/gin"
"github.com/songquanpeng/one-api/common"
"github.com/songquanpeng/one-api/common/ctxkey"
"github.com/songquanpeng/one-api/common/logger"
@@ -209,8 +210,17 @@ func RelayImageHelper(c *gin.Context, relayMode int) *relaymodel.ErrorWithStatus
}
if quota != 0 {
tokenName := c.GetString(ctxkey.TokenName)
logContent := fmt.Sprintf("模型倍率 %.2f，分组倍率 %.2f", modelRatio, groupRatio)
model.RecordConsumeLog(ctx, meta.UserId, meta.ChannelId, 0, 0, imageRequest.Model, tokenName, quota, logContent)
logContent := fmt.Sprintf("倍率：%.2f × %.2f", modelRatio, groupRatio)
model.RecordConsumeLog(ctx, &model.Log{
UserId: meta.UserId,
ChannelId: meta.ChannelId,
PromptTokens: 0,
CompletionTokens: 0,
ModelName: imageRequest.Model,
TokenName: tokenName,
Quota: int(quota),
Content: logContent,
})
model.UpdateUserUsedQuotaAndRequestCount(meta.UserId, quota)
channelId := c.GetInt(ctxkey.ChannelId)
model.UpdateChannelUsedQuota(channelId, quota)

11
relay/controller/text.go
View File

@@ -4,11 +4,12 @@ import (
"bytes"
"encoding/json"
"fmt"
"github.com/songquanpeng/one-api/common/config"
"io"
"net/http"
"github.com/gin-gonic/gin"
"github.com/songquanpeng/one-api/common/config"
"github.com/songquanpeng/one-api/common/logger"
"github.com/songquanpeng/one-api/relay"
"github.com/songquanpeng/one-api/relay/adaptor"
@@ -37,7 +38,7 @@ func RelayTextHelper(c *gin.Context) *model.ErrorWithStatusCode {
textRequest.Model, _ = getMappedModelName(textRequest.Model, meta.ModelMapping)
meta.ActualModelName = textRequest.Model
// set system prompt if not empty
systemPromptReset := setSystemPrompt(ctx, textRequest, meta.SystemPrompt)
systemPromptReset := setSystemPrompt(ctx, textRequest, meta.ForcedSystemPrompt)
// get model ratio & group ratio
modelRatio := billingratio.GetModelRatio(textRequest.Model, meta.ChannelType)
groupRatio := billingratio.GetGroupRatio(meta.Group)
@@ -87,7 +88,11 @@ 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 !config.EnforceIncludeUsage && 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 &&
meta.ForcedSystemPrompt == "" {
// no need to convert request for openai
return c.Request.Body, nil
}

41
relay/meta/relay_meta.go
View File

@@ -1,12 +1,15 @@
package meta
import (
"strings"
"time"
"github.com/gin-gonic/gin"
"github.com/songquanpeng/one-api/common/ctxkey"
"github.com/songquanpeng/one-api/model"
"github.com/songquanpeng/one-api/relay/channeltype"
"github.com/songquanpeng/one-api/relay/relaymode"
"strings"
)
type Meta struct {
@@ -27,27 +30,29 @@ type Meta struct {
// OriginModelName is the model name from the raw user request
OriginModelName string
// ActualModelName is the model name after mapping
ActualModelName string
RequestURLPath string
PromptTokens int // only for DoResponse
SystemPrompt string
ActualModelName string
RequestURLPath string
PromptTokens int // only for DoResponse
ForcedSystemPrompt string
StartTime time.Time
}
func GetByContext(c *gin.Context) *Meta {
meta := Meta{
Mode: relaymode.GetByPath(c.Request.URL.Path),
ChannelType: c.GetInt(ctxkey.Channel),
ChannelId: c.GetInt(ctxkey.ChannelId),
TokenId: c.GetInt(ctxkey.TokenId),
TokenName: c.GetString(ctxkey.TokenName),
UserId: c.GetInt(ctxkey.Id),
Group: c.GetString(ctxkey.Group),
ModelMapping: c.GetStringMapString(ctxkey.ModelMapping),
OriginModelName: c.GetString(ctxkey.RequestModel),
BaseURL: c.GetString(ctxkey.BaseURL),
APIKey: strings.TrimPrefix(c.Request.Header.Get("Authorization"), "Bearer "),
RequestURLPath: c.Request.URL.String(),
SystemPrompt: c.GetString(ctxkey.SystemPrompt),
Mode: relaymode.GetByPath(c.Request.URL.Path),
ChannelType: c.GetInt(ctxkey.Channel),
ChannelId: c.GetInt(ctxkey.ChannelId),
TokenId: c.GetInt(ctxkey.TokenId),
TokenName: c.GetString(ctxkey.TokenName),
UserId: c.GetInt(ctxkey.Id),
Group: c.GetString(ctxkey.Group),
ModelMapping: c.GetStringMapString(ctxkey.ModelMapping),
OriginModelName: c.GetString(ctxkey.RequestModel),
BaseURL: c.GetString(ctxkey.BaseURL),
APIKey: strings.TrimPrefix(c.Request.Header.Get("Authorization"), "Bearer "),
RequestURLPath: c.Request.URL.String(),
ForcedSystemPrompt: c.GetString(ctxkey.SystemPrompt),
StartTime: time.Now(),
}
cfg, ok := c.Get(ctxkey.Config)
if ok {

1
relay/model/general.go
View File

@@ -26,6 +26,7 @@ type GeneralOpenAIRequest struct {
Messages []Message `json:"messages,omitempty"`
Model string `json:"model,omitempty"`
Store *bool `json:"store,omitempty"`
ReasoningEffort *string `json:"reasoning_effort,omitempty"`
Metadata any `json:"metadata,omitempty"`
FrequencyPenalty *float64 `json:"frequency_penalty,omitempty"`
LogitBias any `json:"logit_bias,omitempty"`

11
relay/model/message.go
View File

@@ -3,11 +3,12 @@ package model
import "encoding/json"
type Message struct {
Role string `json:"role,omitempty"`
Content any `json:"content,omitempty"`
Name *string `json:"name,omitempty"`
ToolCalls []Tool `json:"tool_calls,omitempty"`
ToolCallId string `json:"tool_call_id,omitempty"`
Role string `json:"role,omitempty"`
Content any `json:"content,omitempty"`
ReasoningContent any `json:"reasoning_content,omitempty"`
Name *string `json:"name,omitempty"`
ToolCalls []Tool `json:"tool_calls,omitempty"`
ToolCallId string `json:"tool_call_id,omitempty"`
}
func (m Message) IsStringContent() bool {

8
relay/model/misc.go
View File

@@ -4,6 +4,14 @@ type Usage struct {
PromptTokens int `json:"prompt_tokens"`
CompletionTokens int `json:"completion_tokens"`
TotalTokens int `json:"total_tokens"`
CompletionTokensDetails *CompletionTokensDetails `json:"completion_tokens_details,omitempty"`
}
type CompletionTokensDetails struct {
ReasoningTokens int `json:"reasoning_tokens"`
AcceptedPredictionTokens int `json:"accepted_prediction_tokens"`
RejectedPredictionTokens int `json:"rejected_prediction_tokens"`
}
type Error struct {