From 0ad609ade6dad16bb9ec45421083c73b64ede7d1 Mon Sep 17 00:00:00 2001
From: WqyJh <781345688@qq.com>
Date: Tue, 14 Jan 2025 14:38:26 +0800
Subject: [PATCH 1/6] refactor: add GetRatio to Adaptor
---
controller/model.go | 4 +-
model/option.go | 14 +-
relay/adaptor/ai360/constants.go | 12 +-
relay/adaptor/aiproxy/adaptor.go | 16 +-
relay/adaptor/aiproxy/constants.go | 6 +-
relay/adaptor/ali/adaptor.go | 12 +-
relay/adaptor/ali/constants.go | 152 +++++++++++++++---
relay/adaptor/ali/main.go | 11 +-
relay/adaptor/anthropic/adaptor.go | 7 +-
relay/adaptor/anthropic/constants.go | 23 +--
relay/adaptor/aws/adaptor.go | 24 ++-
relay/adaptor/aws/claude/adapter.go | 6 +
relay/adaptor/aws/claude/main.go | 2 +
relay/adaptor/aws/llama3/adapter.go | 6 +
relay/adaptor/aws/llama3/main.go | 6 +
relay/adaptor/aws/utils/adaptor.go | 2 +
relay/adaptor/baichuan/constants.go | 11 +-
relay/adaptor/baidu/adaptor.go | 11 +-
relay/adaptor/baidu/constants.go | 52 ++++--
relay/adaptor/cloudflare/adaptor.go | 5 +
relay/adaptor/cohere/adaptor.go | 7 +-
relay/adaptor/cohere/constant.go | 25 +--
relay/adaptor/common.go | 22 +++
relay/adaptor/coze/adaptor.go | 12 +-
relay/adaptor/coze/constants.go | 4 +-
relay/adaptor/deepl/adaptor.go | 16 +-
relay/adaptor/deepl/constants.go | 11 +-
relay/adaptor/deepseek/constants.go | 8 +-
relay/adaptor/doubao/constants.go | 20 +--
relay/adaptor/gemini/adaptor.go | 7 +-
relay/adaptor/gemini/constants.go | 54 ++++++-
relay/adaptor/groq/constants.go | 37 ++---
relay/adaptor/interface.go | 2 +
relay/adaptor/lingyiwanwu/constants.go | 10 +-
relay/adaptor/minimax/constants.go | 18 ++-
relay/adaptor/mistral/constants.go | 24 ++-
relay/adaptor/moonshot/constants.go | 11 +-
relay/adaptor/novita/constants.go | 47 ++++--
relay/adaptor/ollama/adaptor.go | 12 +-
relay/adaptor/ollama/constants.go | 18 ++-
relay/adaptor/openai/adaptor.go | 10 +-
relay/adaptor/openai/compatible.go | 35 ++--
relay/adaptor/openai/constants.go | 82 +++++++---
relay/adaptor/palm/adaptor.go | 12 +-
relay/adaptor/palm/constants.go | 6 +-
relay/adaptor/proxy/adaptor.go | 5 +
relay/adaptor/replicate/adaptor.go | 10 +-
relay/adaptor/replicate/constant.go | 99 ++++++------
relay/adaptor/siliconflow/constants.go | 68 ++++----
relay/adaptor/stepfun/constants.go | 24 +--
relay/adaptor/tencent/adaptor.go | 20 ++-
relay/adaptor/tencent/constants.go | 22 ++-
relay/adaptor/togetherai/constants.go | 14 +-
relay/adaptor/vertexai/adaptor.go | 26 ++-
relay/adaptor/vertexai/claude/adapter.go | 22 ++-
relay/adaptor/vertexai/gemini/adapter.go | 16 +-
relay/adaptor/vertexai/registry.go | 9 +-
relay/adaptor/xai/constants.go | 6 +-
relay/adaptor/xunfei/adaptor.go | 18 ++-
relay/adaptor/xunfei/constants.go | 20 +--
relay/adaptor/zhipu/adaptor.go | 14 +-
relay/adaptor/zhipu/constants.go | 25 ++-
relay/adaptor/zhipu/main.go | 12 +-
relay/billing/billing.go | 15 +-
relay/billing/ratio/model.go | 52 +++++-
relay/controller/audio.go | 14 +-
relay/controller/helper.go | 38 ++++-
relay/controller/image.go | 7 +-
relay/controller/text.go | 22 +--
.../src/components/OperationSetting.js | 35 +++-
70 files changed, 1038 insertions(+), 467 deletions(-)
diff --git a/controller/model.go b/controller/model.go
index dcbe709e..b9bf6281 100644
--- a/controller/model.go
+++ b/controller/model.go
@@ -6,6 +6,7 @@ import (
"github.com/songquanpeng/one-api/common/ctxkey"
"github.com/songquanpeng/one-api/model"
relay "github.com/songquanpeng/one-api/relay"
+ "github.com/songquanpeng/one-api/relay/adaptor"
"github.com/songquanpeng/one-api/relay/adaptor/openai"
"github.com/songquanpeng/one-api/relay/apitype"
"github.com/songquanpeng/one-api/relay/channeltype"
@@ -86,7 +87,8 @@ func init() {
if channelType == channeltype.Azure {
continue
}
- channelName, channelModelList := openai.GetCompatibleChannelMeta(channelType)
+ channelName, channelRatioMap := openai.GetCompatibleChannelMeta(channelType)
+ channelModelList := adaptor.GetModelListHelper(channelRatioMap)
for _, modelName := range channelModelList {
models = append(models, OpenAIModels{
Id: modelName,
diff --git a/model/option.go b/model/option.go
index 8fd30aee..c749b231 100644
--- a/model/option.go
+++ b/model/option.go
@@ -1,12 +1,13 @@
package model
import (
- "github.com/songquanpeng/one-api/common/config"
- "github.com/songquanpeng/one-api/common/logger"
- billingratio "github.com/songquanpeng/one-api/relay/billing/ratio"
"strconv"
"strings"
"time"
+
+ "github.com/songquanpeng/one-api/common/config"
+ "github.com/songquanpeng/one-api/common/logger"
+ billingratio "github.com/songquanpeng/one-api/relay/billing/ratio"
)
type Option struct {
@@ -70,6 +71,7 @@ func InitOptionMap() {
config.OptionMap["ModelRatio"] = billingratio.ModelRatio2JSONString()
config.OptionMap["GroupRatio"] = billingratio.GroupRatio2JSONString()
config.OptionMap["CompletionRatio"] = billingratio.CompletionRatio2JSONString()
+ config.OptionMap["Ratio"] = billingratio.Ratio2JSONString()
config.OptionMap["TopUpLink"] = config.TopUpLink
config.OptionMap["ChatLink"] = config.ChatLink
config.OptionMap["QuotaPerUnit"] = strconv.FormatFloat(config.QuotaPerUnit, 'f', -1, 64)
@@ -223,12 +225,14 @@ func updateOptionMap(key string, value string) (err error) {
config.PreConsumedQuota, _ = strconv.ParseInt(value, 10, 64)
case "RetryTimes":
config.RetryTimes, _ = strconv.Atoi(value)
- case "ModelRatio":
+ case "ModelRatio": // Deprecated
err = billingratio.UpdateModelRatioByJSONString(value)
case "GroupRatio":
err = billingratio.UpdateGroupRatioByJSONString(value)
- case "CompletionRatio":
+ case "CompletionRatio": // Deprecated
err = billingratio.UpdateCompletionRatioByJSONString(value)
+ case "Ratio":
+ err = billingratio.UpdateRatioByJSONString(value)
case "TopUpLink":
config.TopUpLink = value
case "ChatLink":
diff --git a/relay/adaptor/ai360/constants.go b/relay/adaptor/ai360/constants.go
index cfc3cb28..dd3e6e5e 100644
--- a/relay/adaptor/ai360/constants.go
+++ b/relay/adaptor/ai360/constants.go
@@ -1,8 +1,10 @@
package ai360
-var ModelList = []string{
- "360GPT_S2_V9",
- "embedding-bert-512-v1",
- "embedding_s1_v1",
- "semantic_similarity_s1_v1",
+import "github.com/songquanpeng/one-api/relay/billing/ratio"
+
+var RatioMap = map[string]ratio.Ratio{
+ "360GPT_S2_V9": {Input: 0.012 * ratio.RMB, Output: 0.012 * ratio.RMB},
+ "embedding-bert-512-v1": {Input: 0.0001 * ratio.RMB, Output: 0},
+ "embedding_s1_v1": {Input: 0.0001 * ratio.RMB, Output: 0},
+ "semantic_similarity_s1_v1": {Input: 0.0001 * ratio.RMB, Output: 0},
}
diff --git a/relay/adaptor/aiproxy/adaptor.go b/relay/adaptor/aiproxy/adaptor.go
index 42d49c0a..30317f3a 100644
--- a/relay/adaptor/aiproxy/adaptor.go
+++ b/relay/adaptor/aiproxy/adaptor.go
@@ -3,12 +3,14 @@ package aiproxy
import (
"errors"
"fmt"
- "github.com/gin-gonic/gin"
- "github.com/songquanpeng/one-api/relay/adaptor"
- "github.com/songquanpeng/one-api/relay/meta"
- "github.com/songquanpeng/one-api/relay/model"
"io"
"net/http"
+
+ "github.com/gin-gonic/gin"
+ "github.com/songquanpeng/one-api/relay/adaptor"
+ "github.com/songquanpeng/one-api/relay/billing/ratio"
+ "github.com/songquanpeng/one-api/relay/meta"
+ "github.com/songquanpeng/one-api/relay/model"
)
type Adaptor struct {
@@ -58,8 +60,12 @@ func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Met
return
}
+func (a *Adaptor) GetRatio(meta *meta.Meta) *ratio.Ratio {
+ return adaptor.GetRatioHelper(meta, RatioMap)
+}
+
func (a *Adaptor) GetModelList() []string {
- return ModelList
+ return adaptor.GetModelListHelper(RatioMap)
}
func (a *Adaptor) GetChannelName() string {
diff --git a/relay/adaptor/aiproxy/constants.go b/relay/adaptor/aiproxy/constants.go
index 818d2709..aa684c90 100644
--- a/relay/adaptor/aiproxy/constants.go
+++ b/relay/adaptor/aiproxy/constants.go
@@ -2,8 +2,4 @@ package aiproxy
import "github.com/songquanpeng/one-api/relay/adaptor/openai"
-var ModelList = []string{""}
-
-func init() {
- ModelList = openai.ModelList
-}
+var RatioMap = openai.RatioMap
diff --git a/relay/adaptor/ali/adaptor.go b/relay/adaptor/ali/adaptor.go
index 4aa8a11a..f8ee1d78 100644
--- a/relay/adaptor/ali/adaptor.go
+++ b/relay/adaptor/ali/adaptor.go
@@ -3,13 +3,15 @@ package ali
import (
"errors"
"fmt"
+ "io"
+ "net/http"
+
"github.com/gin-gonic/gin"
"github.com/songquanpeng/one-api/relay/adaptor"
+ "github.com/songquanpeng/one-api/relay/billing/ratio"
"github.com/songquanpeng/one-api/relay/meta"
"github.com/songquanpeng/one-api/relay/model"
"github.com/songquanpeng/one-api/relay/relaymode"
- "io"
- "net/http"
)
// https://help.aliyun.com/zh/dashscope/developer-reference/api-details
@@ -96,8 +98,12 @@ func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Met
return
}
+func (a *Adaptor) GetRatio(meta *meta.Meta) *ratio.Ratio {
+ return adaptor.GetRatioHelper(meta, RatioMap)
+}
+
func (a *Adaptor) GetModelList() []string {
- return ModelList
+ return adaptor.GetModelListHelper(RatioMap)
}
func (a *Adaptor) GetChannelName() string {
diff --git a/relay/adaptor/ali/constants.go b/relay/adaptor/ali/constants.go
index f3d99520..ab4be887 100644
--- a/relay/adaptor/ali/constants.go
+++ b/relay/adaptor/ali/constants.go
@@ -1,23 +1,135 @@
package ali
-var ModelList = []string{
- "qwen-turbo", "qwen-turbo-latest",
- "qwen-plus", "qwen-plus-latest",
- "qwen-max", "qwen-max-latest",
- "qwen-max-longcontext",
- "qwen-vl-max", "qwen-vl-max-latest", "qwen-vl-plus", "qwen-vl-plus-latest",
- "qwen-vl-ocr", "qwen-vl-ocr-latest",
- "qwen-audio-turbo",
- "qwen-math-plus", "qwen-math-plus-latest", "qwen-math-turbo", "qwen-math-turbo-latest",
- "qwen-coder-plus", "qwen-coder-plus-latest", "qwen-coder-turbo", "qwen-coder-turbo-latest",
- "qwq-32b-preview", "qwen2.5-72b-instruct", "qwen2.5-32b-instruct", "qwen2.5-14b-instruct", "qwen2.5-7b-instruct", "qwen2.5-3b-instruct", "qwen2.5-1.5b-instruct", "qwen2.5-0.5b-instruct",
- "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",
- "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",
+import "github.com/songquanpeng/one-api/relay/billing/ratio"
+
+// https://help.aliyun.com/zh/model-studio/getting-started/models
+// https://help.aliyun.com/zh/dashscope/developer-reference/tongyi-thousand-questions-metering-and-billing
+var RatioMap = map[string]ratio.Ratio{
+ "qwen-long": {Input: 0.0005 * ratio.RMB, Output: 0.002 * ratio.RMB},
+ "qwen-turbo": {Input: 0.0003 * ratio.RMB, Output: 0.0006 * ratio.RMB},
+ "qwen-turbo-latest": {Input: 0.0003 * ratio.RMB, Output: 0.0006 * ratio.RMB},
+ "qwen-turbo-2024-09-19": {Input: 0.0003 * ratio.RMB, Output: 0.0006 * ratio.RMB},
+ "qwen-turbo-0919": {Input: 0.0003 * ratio.RMB, Output: 0.0006 * ratio.RMB},
+ "qwen-turbo-2024-06-24": {Input: 0.0003 * ratio.RMB, Output: 0.0006 * ratio.RMB},
+ "qwen-turbo-0624": {Input: 0.0003 * ratio.RMB, Output: 0.0006 * ratio.RMB},
+ "qwen-turbo-2024-02-06": {Input: 0.002 * ratio.RMB, Output: 0.006 * ratio.RMB},
+ "qwen-turbo-0206": {Input: 0.002 * ratio.RMB, Output: 0.006 * ratio.RMB},
+ "qwen-plus": {Input: 0.0008 * ratio.RMB, Output: 0.002 * ratio.RMB},
+ "qwen-plus-latest": {Input: 0.0008 * ratio.RMB, Output: 0.002 * ratio.RMB},
+ "qwen-plus-2024-09-19": {Input: 0.0008 * ratio.RMB, Output: 0.002 * ratio.RMB},
+ "qwen-plus-0919": {Input: 0.0008 * ratio.RMB, Output: 0.002 * ratio.RMB},
+ "qwen-plus-2024-08-06": {Input: 0.004 * ratio.RMB, Output: 0.012 * ratio.RMB},
+ "qwen-plus-0806": {Input: 0.004 * ratio.RMB, Output: 0.012 * ratio.RMB},
+ "qwen-plus-2024-07-23": {Input: 0.004 * ratio.RMB, Output: 0.012 * ratio.RMB},
+ "qwen-plus-0723": {Input: 0.004 * ratio.RMB, Output: 0.012 * ratio.RMB},
+ "qwen-plus-2024-06-24": {Input: 0.004 * ratio.RMB, Output: 0.012 * ratio.RMB},
+ "qwen-plus-0624": {Input: 0.004 * ratio.RMB, Output: 0.012 * ratio.RMB},
+ "qwen-plus-2024-02-06": {Input: 0.004 * ratio.RMB, Output: 0.012 * ratio.RMB},
+ "qwen-plus-0206": {Input: 0.004 * ratio.RMB, Output: 0.012 * ratio.RMB},
+ "qwen-max": {Input: 0.02 * ratio.RMB, Output: 0.06 * ratio.RMB},
+ "qwen-max-latest": {Input: 0.02 * ratio.RMB, Output: 0.06 * ratio.RMB},
+ "qwen-max-2024-09-19": {Input: 0.02 * ratio.RMB, Output: 0.06 * ratio.RMB},
+ "qwen-max-0919": {Input: 0.02 * ratio.RMB, Output: 0.06 * ratio.RMB},
+ "qwen-max-2024-04-28": {Input: 0.04 * ratio.RMB, Output: 0.12 * ratio.RMB},
+ "qwen-max-0428": {Input: 0.04 * ratio.RMB, Output: 0.12 * ratio.RMB},
+ "qwen-max-2024-04-03": {Input: 0.04 * ratio.RMB, Output: 0.12 * ratio.RMB},
+ "qwen-max-0403": {Input: 0.04 * ratio.RMB, Output: 0.12 * ratio.RMB},
+ "qwen-max-2024-01-07": {Input: 0.04 * ratio.RMB, Output: 0.12 * ratio.RMB},
+ "qwen-max-0107": {Input: 0.04 * ratio.RMB, Output: 0.12 * ratio.RMB},
+ "qwen-vl-max": {Input: 0.003 * ratio.RMB, Output: 0.009 * ratio.RMB},
+ "qwen-vl-max-latest": {Input: 0.003 * ratio.RMB, Output: 0.009 * ratio.RMB},
+ "qwen-vl-max-2024-12-30": {Input: 0.003 * ratio.RMB, Output: 0.009 * ratio.RMB},
+ "qwen-vl-max-1230": {Input: 0.003 * ratio.RMB, Output: 0.009 * ratio.RMB},
+ "qwen-vl-max-2024-11-19": {Input: 0.003 * ratio.RMB, Output: 0.009 * ratio.RMB},
+ "qwen-vl-max-1119": {Input: 0.003 * ratio.RMB, Output: 0.009 * ratio.RMB},
+ "qwen-vl-max-2024-10-30": {Input: 0.02 * ratio.RMB, Output: 0.02 * ratio.RMB},
+ "qwen-vl-max-1030": {Input: 0.02 * ratio.RMB, Output: 0.02 * ratio.RMB},
+ "qwen-vl-max-2024-08-09": {Input: 0.02 * ratio.RMB, Output: 0.02 * ratio.RMB},
+ "qwen-vl-max-0809": {Input: 0.02 * ratio.RMB, Output: 0.02 * ratio.RMB},
+ "qwen-vl-max-2024-02-01": {Input: 0.02 * ratio.RMB, Output: 0.02 * ratio.RMB},
+ "qwen-vl-max-0201": {Input: 0.02 * ratio.RMB, Output: 0.02 * ratio.RMB},
+ "qwen-vl-plus": {Input: 0.0015 * ratio.RMB, Output: 0.0045 * ratio.RMB},
+ "qwen-vl-plus-latest": {Input: 0.0015 * ratio.RMB, Output: 0.0045 * ratio.RMB},
+ "qwen-vl-plus-2024-08-09": {Input: 0.0015 * ratio.RMB, Output: 0.0045 * ratio.RMB},
+ "qwen-vl-plus-0809": {Input: 0.0015 * ratio.RMB, Output: 0.0045 * ratio.RMB},
+ "qwen-vl-plus-2023-12-01": {Input: 0.008 * ratio.RMB, Output: 0.008 * ratio.RMB},
+ "qwen-vl-ocr": {Input: 0.005 * ratio.RMB, Output: 0.005 * ratio.RMB},
+ "qwen-vl-ocr-latest": {Input: 0.005 * ratio.RMB, Output: 0.005 * ratio.RMB},
+ "qwen-vl-ocr-2024-10-28": {Input: 0.005 * ratio.RMB, Output: 0.005 * ratio.RMB},
+ "qwen-audio-turbo": {Input: 0.1, Output: 0.1}, // 目前仅供免费体验。免费额度用完后不可调用，敬请关注后续动态。
+ "qwen-audio-turbo-latest": {Input: 0.1, Output: 0.1}, // 目前仅供免费体验。免费额度用完后不可调用，敬请关注后续动态。
+ "qwen-audio-turbo-2024-12-04": {Input: 0.1, Output: 0.1}, // 目前仅供免费体验。免费额度用完后不可调用，敬请关注后续动态。
+ "qwen-audio-turbo-1204": {Input: 0.1, Output: 0.1}, // 目前仅供免费体验。免费额度用完后不可调用，敬请关注后续动态。
+ "qwen-audio-turbo-2024-08-07": {Input: 0.1, Output: 0.1}, // 目前仅供免费体验。免费额度用完后不可调用，敬请关注后续动态。
+ "qwen-audio-turbo-0807": {Input: 0.1, Output: 0.1}, // 目前仅供免费体验。免费额度用完后不可调用，敬请关注后续动态。
+ "qwen-math-plus": {Input: 0.004 * ratio.RMB, Output: 0.012 * ratio.RMB},
+ "qwen-math-plus-latest": {Input: 0.004 * ratio.RMB, Output: 0.012 * ratio.RMB},
+ "qwen-math-plus-2024-09-19": {Input: 0.004 * ratio.RMB, Output: 0.012 * ratio.RMB},
+ "qwen-math-plus-0919": {Input: 0.004 * ratio.RMB, Output: 0.012 * ratio.RMB},
+ "qwen-math-plus-2024-08-16": {Input: 0.004 * ratio.RMB, Output: 0.012 * ratio.RMB},
+ "qwen-math-plus-0816": {Input: 0.004 * ratio.RMB, Output: 0.012 * ratio.RMB},
+ "qwen-math-turbo": {Input: 0.004 * ratio.RMB, Output: 0.012 * ratio.RMB},
+ "qwen-math-turbo-latest": {Input: 0.004 * ratio.RMB, Output: 0.012 * ratio.RMB},
+ "qwen-math-turbo-2024-09-19": {Input: 0.004 * ratio.RMB, Output: 0.012 * ratio.RMB},
+ "qwen-math-turbo-0919": {Input: 0.004 * ratio.RMB, Output: 0.012 * ratio.RMB},
+ "qwen-coder-plus": {Input: 0.0035 * ratio.RMB, Output: 0.007 * ratio.RMB},
+ "qwen-coder-plus-latest": {Input: 0.0035 * ratio.RMB, Output: 0.007 * ratio.RMB},
+ "qwen-coder-plus-2024-11-06": {Input: 0.0035 * ratio.RMB, Output: 0.007 * ratio.RMB},
+ "qwen-coder-plus-1106": {Input: 0.0035 * ratio.RMB, Output: 0.007 * ratio.RMB},
+ "qwen-coder-turbo": {Input: 0.002 * ratio.RMB, Output: 0.006 * ratio.RMB},
+ "qwen-coder-turbo-latest": {Input: 0.002 * ratio.RMB, Output: 0.006 * ratio.RMB},
+ "qwen-coder-turbo-2024-09-19": {Input: 0.002 * ratio.RMB, Output: 0.006 * ratio.RMB},
+ "qwen-coder-turbo-0919": {Input: 0.002 * ratio.RMB, Output: 0.006 * ratio.RMB},
+ "qwq-32b-preview": {Input: 0.0035 * ratio.RMB, Output: 0.007 * ratio.RMB},
+ "qwen2.5-72b-instruct": {Input: 0.004 * ratio.RMB, Output: 0.012 * ratio.RMB},
+ "qwen2.5-32b-instruct": {Input: 0.0035 * ratio.RMB, Output: 0.007 * ratio.RMB},
+ "qwen2.5-14b-instruct": {Input: 0.002 * ratio.RMB, Output: 0.006 * ratio.RMB},
+ "qwen2.5-7b-instruct": {Input: 0.001 * ratio.RMB, Output: 0.002 * ratio.RMB},
+ "qwen2.5-3b-instruct": {Input: 0.1, Output: 0.1}, // 限时免费
+ "qwen2.5-1.5b-instruct": {Input: 0.1, Output: 0.1}, // 限时免费
+ "qwen2.5-0.5b-instruct": {Input: 0.1, Output: 0.1}, // 限时免费
+ "qwen2-72b-instruct": {Input: 0.004 * ratio.RMB, Output: 0.012 * ratio.RMB},
+ "qwen2-57b-a14b-instruct": {Input: 0.0035 * ratio.RMB, Output: 0.007 * ratio.RMB},
+ "qwen2-7b-instruct": {Input: 0.001 * ratio.RMB, Output: 0.002 * ratio.RMB},
+ "qwen2-1.5b-instruct": {Input: 0.1, Output: 0.1}, // 限时免费
+ "qwen2-0.5b-instruct": {Input: 0.1, Output: 0.1}, // 限时免费
+ "qwen1.5-110b-chat": {Input: 0.007 * ratio.RMB, Output: 0.014 * ratio.RMB},
+ "qwen1.5-72b-chat": {Input: 0.005 * ratio.RMB, Output: 0.01 * ratio.RMB},
+ "qwen1.5-32b-chat": {Input: 0.0035 * ratio.RMB, Output: 0.007 * ratio.RMB},
+ "qwen1.5-14b-chat": {Input: 0.002 * ratio.RMB, Output: 0.004 * ratio.RMB},
+ "qwen1.5-7b-chat": {Input: 0.001 * ratio.RMB, Output: 0.002 * ratio.RMB},
+ "qwen1.5-1.8b-chat": {Input: 0.1, Output: 0.1}, // 限时免费
+ "qwen1.5-0.5b-chat": {Input: 0.1, Output: 0.1}, // 限时免费
+ "qwen-72b-chat": {Input: 0.02 * ratio.RMB, Output: 0.02 * ratio.RMB},
+ "qwen-14b-chat": {Input: 0.008 * ratio.RMB, Output: 0.008 * ratio.RMB},
+ "qwen-7b-chat": {Input: 0.006 * ratio.RMB, Output: 0.006 * ratio.RMB},
+ "qwen-1.8b-chat": {Input: 0.1, Output: 0.1}, // 限时免费
+ "qwen-1.8b-longcontext-chat": {Input: 0.1, Output: 0.1}, // 限时免费（需申请）
+ "qwen2-vl-72b-instruct": {Input: 0.1, Output: 0.1}, // 目前仅供免费体验。免费额度用完后不可调用，敬请关注后续动态。
+ "qwen2-vl-7b-instruct": {Input: 0.1, Output: 0.1}, // 目前仅供免费体验。免费额度用完后不可调用，敬请关注后续动态。
+ "qwen2-vl-2b-instruct": {Input: 0.1, Output: 0.1}, // 限时免费
+ "qwen-vl-v1": {Input: 0.1, Output: 0.1}, // 目前仅供免费体验。免费额度用完后不可调用，敬请关注后续动态。
+ "qwen-vl-chat-v1": {Input: 0.1, Output: 0.1}, // 目前仅供免费体验。免费额度用完后不可调用，敬请关注后续动态。
+ "qwen2-audio-instruct": {Input: 0.1, Output: 0.1}, // 目前仅供免费体验。免费额度用完后不可调用，敬请关注后续动态。
+ "qwen-audio-chat": {Input: 0.1, Output: 0.1}, // 目前仅供免费体验。免费额度用完后不可调用，敬请关注后续动态。
+ "qwen2.5-math-72b-instruct": {Input: 0.004 * ratio.RMB, Output: 0.012 * ratio.RMB},
+ "qwen2.5-math-7b-instruct": {Input: 0.001 * ratio.RMB, Output: 0.002 * ratio.RMB},
+ "qwen2.5-math-1.5b-instruct": {Input: 0.1, Output: 0.1}, // 限时免费
+ "qwen2-math-72b-instruct": {Input: 0.004 * ratio.RMB, Output: 0.012 * ratio.RMB},
+ "qwen2-math-7b-instruct": {Input: 0.001 * ratio.RMB, Output: 0.002 * ratio.RMB},
+ "qwen2-math-1.5b-instruct": {Input: 0.1, Output: 0.1}, // 目前仅供免费体验。免费额度用完后不可调用，敬请关注后续动态。
+ "qwen2.5-coder-32b-instruct": {Input: 0.0035 * ratio.RMB, Output: 0.007 * ratio.RMB},
+ "qwen2.5-coder-14b-instruct": {Input: 0.002 * ratio.RMB, Output: 0.006 * ratio.RMB},
+ "qwen2.5-coder-7b-instruct": {Input: 0.001 * ratio.RMB, Output: 0.002 * ratio.RMB},
+ "qwen2.5-coder-3b-instruct": {Input: 0.1, Output: 0.1}, // 限时免费
+ "qwen2.5-coder-1.5b-instruct": {Input: 0.1, Output: 0.1}, // 目前仅供免费体验。免费额度用完后不可调用，敬请关注后续动态。
+ "qwen2.5-coder-0.5b-instruct": {Input: 0.1, Output: 0.1}, // 限时免费
+ "text-embedding-v3": {Input: 0.0007 * ratio.RMB, Output: 0},
+ "text-embedding-v2": {Input: 0.0007 * ratio.RMB, Output: 0},
+ "text-embedding-v1": {Input: 0.0007 * ratio.RMB, Output: 0},
+ "text-embedding-async-v2": {Input: 0.0007 * ratio.RMB, Output: 0},
+ "text-embedding-async-v1": {Input: 0.0007 * ratio.RMB, Output: 0},
+ "ali-stable-diffusion-xl": {Input: 8.00 * ratio.RMB, Output: 0},
+ "ali-stable-diffusion-v1.5": {Input: 8.00 * ratio.RMB, Output: 0},
+ "wanx-v1": {Input: 8.00 * ratio.RMB, Output: 0},
}
diff --git a/relay/adaptor/ali/main.go b/relay/adaptor/ali/main.go
index 6a73c707..c67c150a 100644
--- a/relay/adaptor/ali/main.go
+++ b/relay/adaptor/ali/main.go
@@ -3,12 +3,13 @@ package ali
import (
"bufio"
"encoding/json"
- "github.com/songquanpeng/one-api/common/ctxkey"
- "github.com/songquanpeng/one-api/common/render"
"io"
"net/http"
"strings"
+ "github.com/songquanpeng/one-api/common/ctxkey"
+ "github.com/songquanpeng/one-api/common/render"
+
"github.com/gin-gonic/gin"
"github.com/songquanpeng/one-api/common"
"github.com/songquanpeng/one-api/common/helper"
@@ -119,7 +120,11 @@ func embeddingResponseAli2OpenAI(response *EmbeddingResponse) *openai.EmbeddingR
Object: "list",
Data: make([]openai.EmbeddingResponseItem, 0, len(response.Output.Embeddings)),
Model: "text-embedding-v1",
- Usage: model.Usage{TotalTokens: response.Usage.TotalTokens},
+ Usage: model.Usage{
+ PromptTokens: response.Usage.InputTokens,
+ CompletionTokens: response.Usage.OutputTokens,
+ TotalTokens: response.Usage.TotalTokens,
+ },
}
for _, item := range response.Output.Embeddings {
diff --git a/relay/adaptor/anthropic/adaptor.go b/relay/adaptor/anthropic/adaptor.go
index bd0949be..6cef3571 100644
--- a/relay/adaptor/anthropic/adaptor.go
+++ b/relay/adaptor/anthropic/adaptor.go
@@ -9,6 +9,7 @@ import (
"github.com/gin-gonic/gin"
"github.com/songquanpeng/one-api/relay/adaptor"
+ "github.com/songquanpeng/one-api/relay/billing/ratio"
"github.com/songquanpeng/one-api/relay/meta"
"github.com/songquanpeng/one-api/relay/model"
)
@@ -70,8 +71,12 @@ func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Met
return
}
+func (a *Adaptor) GetRatio(meta *meta.Meta) *ratio.Ratio {
+ return adaptor.GetRatioHelper(meta, RatioMap)
+}
+
func (a *Adaptor) GetModelList() []string {
- return ModelList
+ return adaptor.GetModelListHelper(RatioMap)
}
func (a *Adaptor) GetChannelName() string {
diff --git a/relay/adaptor/anthropic/constants.go b/relay/adaptor/anthropic/constants.go
index 8ea7c4d8..efaab72b 100644
--- a/relay/adaptor/anthropic/constants.go
+++ b/relay/adaptor/anthropic/constants.go
@@ -1,12 +1,17 @@
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",
+import "github.com/songquanpeng/one-api/relay/billing/ratio"
+
+// https://www.anthropic.com/api#pricing
+var RatioMap = map[string]ratio.Ratio{
+ "claude-instant-1.2": {Input: 0.8 * ratio.MILLI_USD, Output: 2.4 * ratio.MILLI_USD},
+ "claude-2.0": {Input: 8.0 * ratio.MILLI_USD, Output: 24 * ratio.MILLI_USD},
+ "claude-2.1": {Input: 8.0 * ratio.MILLI_USD, Output: 24 * ratio.MILLI_USD},
+ "claude-3-haiku-20240307": {Input: 0.25 * ratio.MILLI_USD, Output: 1.25 * ratio.MILLI_USD},
+ "claude-3-5-haiku-20241022": {Input: 0.8 * ratio.MILLI_USD, Output: 4 * ratio.MILLI_USD},
+ "claude-3-sonnet-20240229": {Input: 3 * ratio.MILLI_USD, Output: 15 * ratio.MILLI_USD},
+ "claude-3-opus-20240229": {Input: 15 * ratio.MILLI_USD, Output: 75 * ratio.MILLI_USD},
+ "claude-3-5-sonnet-20240620": {Input: 3 * ratio.MILLI_USD, Output: 15 * ratio.MILLI_USD},
+ "claude-3-5-sonnet-20241022": {Input: 3 * ratio.MILLI_USD, Output: 15 * ratio.MILLI_USD},
+ "claude-3-5-sonnet-latest": {Input: 3 * ratio.MILLI_USD, Output: 15 * ratio.MILLI_USD},
}
diff --git a/relay/adaptor/aws/adaptor.go b/relay/adaptor/aws/adaptor.go
index 62221346..125e0f3f 100644
--- a/relay/adaptor/aws/adaptor.go
+++ b/relay/adaptor/aws/adaptor.go
@@ -10,7 +10,7 @@ import (
"github.com/aws/aws-sdk-go-v2/service/bedrockruntime"
"github.com/gin-gonic/gin"
"github.com/songquanpeng/one-api/relay/adaptor"
- "github.com/songquanpeng/one-api/relay/adaptor/aws/utils"
+ "github.com/songquanpeng/one-api/relay/billing/ratio"
"github.com/songquanpeng/one-api/relay/meta"
"github.com/songquanpeng/one-api/relay/model"
)
@@ -18,8 +18,6 @@ import (
var _ adaptor.Adaptor = new(Adaptor)
type Adaptor struct {
- awsAdapter utils.AwsAdapter
-
Meta *meta.Meta
AwsClient *bedrockruntime.Client
}
@@ -42,15 +40,27 @@ func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.G
return nil, errors.New("adaptor not found")
}
- a.awsAdapter = adaptor
return adaptor.ConvertRequest(c, relayMode, request)
}
func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
- if a.awsAdapter == nil {
- return nil, utils.WrapErr(errors.New("awsAdapter is nil"))
+ adaptor := GetAdaptor(meta.ActualModelName)
+ if adaptor == nil {
+ return nil, &model.ErrorWithStatusCode{
+ StatusCode: http.StatusInternalServerError,
+ Error: model.Error{Message: "adaptor not found"},
+ }
}
- return a.awsAdapter.DoResponse(c, a.AwsClient, meta)
+
+ return adaptor.DoResponse(c, a.AwsClient, meta)
+}
+
+func (a *Adaptor) GetRatio(meta *meta.Meta) *ratio.Ratio {
+ adaptor := GetAdaptor(meta.ActualModelName)
+ if adaptor == nil {
+ return nil
+ }
+ return adaptor.GetRatio(meta)
}
func (a *Adaptor) GetModelList() (models []string) {
diff --git a/relay/adaptor/aws/claude/adapter.go b/relay/adaptor/aws/claude/adapter.go
index eb3c9fb8..1edbecfa 100644
--- a/relay/adaptor/aws/claude/adapter.go
+++ b/relay/adaptor/aws/claude/adapter.go
@@ -5,8 +5,10 @@ import (
"github.com/gin-gonic/gin"
"github.com/pkg/errors"
"github.com/songquanpeng/one-api/common/ctxkey"
+ "github.com/songquanpeng/one-api/relay/adaptor"
"github.com/songquanpeng/one-api/relay/adaptor/anthropic"
"github.com/songquanpeng/one-api/relay/adaptor/aws/utils"
+ "github.com/songquanpeng/one-api/relay/billing/ratio"
"github.com/songquanpeng/one-api/relay/meta"
"github.com/songquanpeng/one-api/relay/model"
)
@@ -35,3 +37,7 @@ func (a *Adaptor) DoResponse(c *gin.Context, awsCli *bedrockruntime.Client, meta
}
return
}
+
+func (a *Adaptor) GetRatio(meta *meta.Meta) *ratio.Ratio {
+ return adaptor.GetRatioHelper(meta, anthropic.RatioMap)
+}
diff --git a/relay/adaptor/aws/claude/main.go b/relay/adaptor/aws/claude/main.go
index 3fe3dfd8..d3b6f059 100644
--- a/relay/adaptor/aws/claude/main.go
+++ b/relay/adaptor/aws/claude/main.go
@@ -38,6 +38,8 @@ var AwsModelIDMap = map[string]string{
"claude-3-5-haiku-20241022": "anthropic.claude-3-5-haiku-20241022-v1:0",
}
+var RatioMap = anthropic.RatioMap
+
func awsModelID(requestModel string) (string, error) {
if awsModelID, ok := AwsModelIDMap[requestModel]; ok {
return awsModelID, nil
diff --git a/relay/adaptor/aws/llama3/adapter.go b/relay/adaptor/aws/llama3/adapter.go
index 83edbc9d..a58b5035 100644
--- a/relay/adaptor/aws/llama3/adapter.go
+++ b/relay/adaptor/aws/llama3/adapter.go
@@ -6,7 +6,9 @@ import (
"github.com/gin-gonic/gin"
"github.com/pkg/errors"
+ "github.com/songquanpeng/one-api/relay/adaptor"
"github.com/songquanpeng/one-api/relay/adaptor/aws/utils"
+ "github.com/songquanpeng/one-api/relay/billing/ratio"
"github.com/songquanpeng/one-api/relay/meta"
"github.com/songquanpeng/one-api/relay/model"
)
@@ -35,3 +37,7 @@ func (a *Adaptor) DoResponse(c *gin.Context, awsCli *bedrockruntime.Client, meta
}
return
}
+
+func (a *Adaptor) GetRatio(meta *meta.Meta) *ratio.Ratio {
+ return adaptor.GetRatioHelper(meta, RatioMap)
+}
diff --git a/relay/adaptor/aws/llama3/main.go b/relay/adaptor/aws/llama3/main.go
index e5fcd89f..a9531ddf 100644
--- a/relay/adaptor/aws/llama3/main.go
+++ b/relay/adaptor/aws/llama3/main.go
@@ -22,6 +22,7 @@ import (
"github.com/songquanpeng/one-api/common/logger"
"github.com/songquanpeng/one-api/relay/adaptor/aws/utils"
"github.com/songquanpeng/one-api/relay/adaptor/openai"
+ "github.com/songquanpeng/one-api/relay/billing/ratio"
relaymodel "github.com/songquanpeng/one-api/relay/model"
)
@@ -32,6 +33,11 @@ var AwsModelIDMap = map[string]string{
"llama3-70b-8192": "meta.llama3-70b-instruct-v1:0",
}
+var RatioMap = map[string]ratio.Ratio{
+ "llama3-8b-8192": {Input: 0.3 * ratio.MILLI_USD, Output: 0.6 * ratio.MILLI_USD},
+ "llama3-70b-8192": {Input: 2.65 * ratio.MILLI_USD, Output: 3.5 * ratio.MILLI_USD},
+}
+
func awsModelID(requestModel string) (string, error) {
if awsModelID, ok := AwsModelIDMap[requestModel]; ok {
return awsModelID, nil
diff --git a/relay/adaptor/aws/utils/adaptor.go b/relay/adaptor/aws/utils/adaptor.go
index 4cb880f2..70072b79 100644
--- a/relay/adaptor/aws/utils/adaptor.go
+++ b/relay/adaptor/aws/utils/adaptor.go
@@ -9,6 +9,7 @@ import (
"github.com/aws/aws-sdk-go-v2/credentials"
"github.com/aws/aws-sdk-go-v2/service/bedrockruntime"
"github.com/gin-gonic/gin"
+ "github.com/songquanpeng/one-api/relay/billing/ratio"
"github.com/songquanpeng/one-api/relay/meta"
"github.com/songquanpeng/one-api/relay/model"
)
@@ -16,6 +17,7 @@ import (
type AwsAdapter interface {
ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error)
DoResponse(c *gin.Context, awsCli *bedrockruntime.Client, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode)
+ GetRatio(meta *meta.Meta) *ratio.Ratio
}
type Adaptor struct {
diff --git a/relay/adaptor/baichuan/constants.go b/relay/adaptor/baichuan/constants.go
index cb20a1ff..4ad41d30 100644
--- a/relay/adaptor/baichuan/constants.go
+++ b/relay/adaptor/baichuan/constants.go
@@ -1,7 +1,10 @@
package baichuan
-var ModelList = []string{
- "Baichuan2-Turbo",
- "Baichuan2-Turbo-192k",
- "Baichuan-Text-Embedding",
+import "github.com/songquanpeng/one-api/relay/billing/ratio"
+
+// https://platform.baichuan-ai.com/price
+var RatioMap = map[string]ratio.Ratio{
+ "Baichuan2-Turbo": {Input: 0.008 * ratio.RMB, Output: 0.008 * ratio.RMB},
+ "Baichuan2-Turbo-192k": {Input: 0.016 * ratio.RMB, Output: 0.016 * ratio.RMB},
+ "Baichuan-Text-Embedding": {Input: 0.001 * ratio.RMB, Output: 0.001 * ratio.RMB},
}
diff --git a/relay/adaptor/baidu/adaptor.go b/relay/adaptor/baidu/adaptor.go
index 15306b95..135313bc 100644
--- a/relay/adaptor/baidu/adaptor.go
+++ b/relay/adaptor/baidu/adaptor.go
@@ -3,15 +3,16 @@ package baidu
import (
"errors"
"fmt"
- "github.com/songquanpeng/one-api/relay/meta"
- "github.com/songquanpeng/one-api/relay/relaymode"
"io"
"net/http"
"strings"
"github.com/gin-gonic/gin"
"github.com/songquanpeng/one-api/relay/adaptor"
+ "github.com/songquanpeng/one-api/relay/billing/ratio"
+ "github.com/songquanpeng/one-api/relay/meta"
"github.com/songquanpeng/one-api/relay/model"
+ "github.com/songquanpeng/one-api/relay/relaymode"
)
type Adaptor struct {
@@ -120,6 +121,10 @@ func (a *Adaptor) DoRequest(c *gin.Context, meta *meta.Meta, requestBody io.Read
return adaptor.DoRequestHelper(a, c, meta, requestBody)
}
+func (a *Adaptor) GetRatio(meta *meta.Meta) *ratio.Ratio {
+ return adaptor.GetRatioHelper(meta, RatioMap)
+}
+
func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
if meta.IsStream {
err, usage = StreamHandler(c, resp)
@@ -135,7 +140,7 @@ func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Met
}
func (a *Adaptor) GetModelList() []string {
- return ModelList
+ return adaptor.GetModelListHelper(RatioMap)
}
func (a *Adaptor) GetChannelName() string {
diff --git a/relay/adaptor/baidu/constants.go b/relay/adaptor/baidu/constants.go
index f952adc6..38083397 100644
--- a/relay/adaptor/baidu/constants.go
+++ b/relay/adaptor/baidu/constants.go
@@ -1,20 +1,38 @@
package baidu
-var ModelList = []string{
- "ERNIE-4.0-8K",
- "ERNIE-3.5-8K",
- "ERNIE-3.5-8K-0205",
- "ERNIE-3.5-8K-1222",
- "ERNIE-Bot-8K",
- "ERNIE-3.5-4K-0205",
- "ERNIE-Speed-8K",
- "ERNIE-Speed-128K",
- "ERNIE-Lite-8K-0922",
- "ERNIE-Lite-8K-0308",
- "ERNIE-Tiny-8K",
- "BLOOMZ-7B",
- "Embedding-V1",
- "bge-large-zh",
- "bge-large-en",
- "tao-8k",
+import "github.com/songquanpeng/one-api/relay/billing/ratio"
+
+// https://cloud.baidu.com/doc/WENXINWORKSHOP/s/hlrk4akp7
+var RatioMap = map[string]ratio.Ratio{
+ "ERNIE-4.0-Turbo-128K": {Input: 0.02 * ratio.RMB, Output: 0.06 * ratio.RMB},
+ "ERNIE-4.0-Turbo-8K": {Input: 0.02 * ratio.RMB, Output: 0.06 * ratio.RMB},
+ "ERNIE-4.0-Turbo-8K-Preview": {Input: 0.02 * ratio.RMB, Output: 0.06 * ratio.RMB},
+ "ERNIE-4.0-Turbo-8K-0628": {Input: 0.02 * ratio.RMB, Output: 0.06 * ratio.RMB},
+ "ERNIE-4.0-8K": {Input: 0.03 * ratio.RMB, Output: 0.09 * ratio.RMB},
+ "ERNIE-4.0-8K-0613": {Input: 0.03 * ratio.RMB, Output: 0.09 * ratio.RMB},
+ "ERNIE-4.0-8K-Latest": {Input: 0.03 * ratio.RMB, Output: 0.09 * ratio.RMB},
+ "ERNIE-4.0-8K-Preview": {Input: 0.03 * ratio.RMB, Output: 0.09 * ratio.RMB},
+ "ERNIE-3.5-128K": {Input: 0.0008 * ratio.RMB, Output: 0.002 * ratio.RMB},
+ "ERNIE-3.5-8K": {Input: 0.0008 * ratio.RMB, Output: 0.002 * ratio.RMB},
+ "ERNIE-3.5-8K-0701": {Input: 0.0008 * ratio.RMB, Output: 0.002 * ratio.RMB},
+ "ERNIE-3.5-8K-Preview": {Input: 0.0008 * ratio.RMB, Output: 0.002 * ratio.RMB},
+ "ERNIE-3.5-8K-0613": {Input: 0.0008 * ratio.RMB, Output: 0.002 * ratio.RMB},
+ "ERNIE-Speed-Pro-128K": {Input: 0.0003 * ratio.RMB, Output: 0.0006 * ratio.RMB},
+ "ERNIE-Novel-8K": {Input: 0.04 * ratio.RMB, Output: 0.12 * ratio.RMB},
+ "ERNIE-Speed-128K": {Input: 0.1, Output: 0.1}, // 免费
+ "ERNIE-Speed-8K": {Input: 0.1, Output: 0.1}, // 免费
+ "ERNIE-Lite-8K": {Input: 0.1, Output: 0.1}, // 免费
+ "ERNIE-Tiny-8K": {Input: 0.1, Output: 0.1}, // 免费
+ "ERNIE-Functions-8K": {Input: 0.004 * ratio.RMB, Output: 0.008 * ratio.RMB},
+ "ERNIE-Character-8K": {Input: 0.0003 * ratio.RMB, Output: 0.0006 * ratio.RMB},
+ "ERNIE-Character-Fiction-8K": {Input: 0.0003 * ratio.RMB, Output: 0.0006 * ratio.RMB},
+ "ERNIE-Character-Fiction-8K-Preview ": {Input: 0.0003 * ratio.RMB, Output: 0.0006 * ratio.RMB},
+ "ERNIE-Lite-Pro-128K": {Input: 0.0002 * ratio.RMB, Output: 0.0004 * ratio.RMB},
+ "Qianfan-Agent-Speed-8K": {Input: 0.0003 * ratio.RMB, Output: 0.0006 * ratio.RMB},
+ "Qianfan-Agent-Lite-8K": {Input: 0.0002 * ratio.RMB, Output: 0.0004 * ratio.RMB},
+ "BLOOMZ-7B": {Input: 0.004 * ratio.RMB, Output: 0.004 * ratio.RMB},
+ "Embedding-V1": {Input: 0.0005 * ratio.RMB, Output: 0},
+ "bge-large-zh": {Input: 0.0005 * ratio.RMB, Output: 0},
+ "bge-large-en": {Input: 0.0005 * ratio.RMB, Output: 0},
+ "tao-8k": {Input: 0.0005 * ratio.RMB, Output: 0},
}
diff --git a/relay/adaptor/cloudflare/adaptor.go b/relay/adaptor/cloudflare/adaptor.go
index 97e3dbb2..4985d5b3 100644
--- a/relay/adaptor/cloudflare/adaptor.go
+++ b/relay/adaptor/cloudflare/adaptor.go
@@ -9,6 +9,7 @@ import (
"github.com/gin-gonic/gin"
"github.com/songquanpeng/one-api/relay/adaptor"
+ "github.com/songquanpeng/one-api/relay/billing/ratio"
"github.com/songquanpeng/one-api/relay/meta"
"github.com/songquanpeng/one-api/relay/model"
"github.com/songquanpeng/one-api/relay/relaymode"
@@ -91,6 +92,10 @@ func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Met
return
}
+func (a *Adaptor) GetRatio(meta *meta.Meta) *ratio.Ratio {
+ return adaptor.GetRatioHelper(meta, nil)
+}
+
func (a *Adaptor) GetModelList() []string {
return ModelList
}
diff --git a/relay/adaptor/cohere/adaptor.go b/relay/adaptor/cohere/adaptor.go
index 6fdb1b04..5d6aa2f4 100644
--- a/relay/adaptor/cohere/adaptor.go
+++ b/relay/adaptor/cohere/adaptor.go
@@ -8,6 +8,7 @@ import (
"github.com/gin-gonic/gin"
"github.com/songquanpeng/one-api/relay/adaptor"
+ "github.com/songquanpeng/one-api/relay/billing/ratio"
"github.com/songquanpeng/one-api/relay/meta"
"github.com/songquanpeng/one-api/relay/model"
)
@@ -55,8 +56,12 @@ func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Met
return
}
+func (a *Adaptor) GetRatio(meta *meta.Meta) *ratio.Ratio {
+ return adaptor.GetRatioHelper(meta, RatioMap)
+}
+
func (a *Adaptor) GetModelList() []string {
- return ModelList
+ return adaptor.GetModelListHelper(RatioMap)
}
func (a *Adaptor) GetChannelName() string {
diff --git a/relay/adaptor/cohere/constant.go b/relay/adaptor/cohere/constant.go
index 9e70652c..be56429d 100644
--- a/relay/adaptor/cohere/constant.go
+++ b/relay/adaptor/cohere/constant.go
@@ -1,14 +1,19 @@
package cohere
-var ModelList = []string{
- "command", "command-nightly",
- "command-light", "command-light-nightly",
- "command-r", "command-r-plus",
-}
+import "github.com/songquanpeng/one-api/relay/billing/ratio"
-func init() {
- num := len(ModelList)
- for i := 0; i < num; i++ {
- ModelList = append(ModelList, ModelList[i]+"-internet")
- }
+// https://cohere.com/pricing
+var RatioMap = map[string]ratio.Ratio{
+ "command": {Input: 1 * ratio.MILLI_USD, Output: 2 * ratio.MILLI_USD},
+ "command-internet": {Input: 1 * ratio.MILLI_USD, Output: 2 * ratio.MILLI_USD},
+ "command-nightly": {Input: 1 * ratio.MILLI_USD, Output: 2 * ratio.MILLI_USD},
+ "command-nightly-internet": {Input: 1 * ratio.MILLI_USD, Output: 2 * ratio.MILLI_USD},
+ "command-light": {Input: 0.3 * ratio.MILLI_USD, Output: 0.6 * ratio.MILLI_USD},
+ "command-light-internet": {Input: 0.3 * ratio.MILLI_USD, Output: 0.6 * ratio.MILLI_USD},
+ "command-light-nightly": {Input: 0.3 * ratio.MILLI_USD, Output: 0.6 * ratio.MILLI_USD},
+ "command-light-nightly-internet": {Input: 0.3 * ratio.MILLI_USD, Output: 0.6 * ratio.MILLI_USD},
+ "command-r": {Input: 0.15 * ratio.MILLI_USD, Output: 0.6 * ratio.MILLI_USD},
+ "command-r-internet": {Input: 0.15 * ratio.MILLI_USD, Output: 0.6 * ratio.MILLI_USD},
+ "command-r-plus": {Input: 2.5 * ratio.MILLI_USD, Output: 10 * ratio.MILLI_USD},
+ "command-r-plus-internet": {Input: 2.5 * ratio.MILLI_USD, Output: 10 * ratio.MILLI_USD},
}
diff --git a/relay/adaptor/common.go b/relay/adaptor/common.go
index 8953d7a3..766cc3b5 100644
--- a/relay/adaptor/common.go
+++ b/relay/adaptor/common.go
@@ -5,6 +5,7 @@ import (
"fmt"
"github.com/gin-gonic/gin"
"github.com/songquanpeng/one-api/common/client"
+ "github.com/songquanpeng/one-api/relay/billing/ratio"
"github.com/songquanpeng/one-api/relay/meta"
"io"
"net/http"
@@ -50,3 +51,24 @@ func DoRequest(c *gin.Context, req *http.Request) (*http.Response, error) {
_ = c.Request.Body.Close()
return resp, nil
}
+
+func GetRatioHelper(meta *meta.Meta, ratioMap map[string]ratio.Ratio) *ratio.Ratio {
+ var result ratio.Ratio
+ if ratio, ok := ratioMap[meta.OriginModelName]; ok {
+ result = ratio
+ return &result
+ }
+ if ratio, ok := ratioMap[meta.ActualModelName]; ok {
+ result = ratio
+ return &result
+ }
+ return nil
+}
+
+func GetModelListHelper(ratioMap map[string]ratio.Ratio) []string {
+ var modelList []string
+ for model := range ratioMap {
+ modelList = append(modelList, model)
+ }
+ return modelList
+}
diff --git a/relay/adaptor/coze/adaptor.go b/relay/adaptor/coze/adaptor.go
index 44f560e8..51d25531 100644
--- a/relay/adaptor/coze/adaptor.go
+++ b/relay/adaptor/coze/adaptor.go
@@ -3,13 +3,15 @@ package coze
import (
"errors"
"fmt"
+ "io"
+ "net/http"
+
"github.com/gin-gonic/gin"
"github.com/songquanpeng/one-api/relay/adaptor"
"github.com/songquanpeng/one-api/relay/adaptor/openai"
+ "github.com/songquanpeng/one-api/relay/billing/ratio"
"github.com/songquanpeng/one-api/relay/meta"
"github.com/songquanpeng/one-api/relay/model"
- "io"
- "net/http"
)
type Adaptor struct {
@@ -66,8 +68,12 @@ func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Met
return
}
+func (a *Adaptor) GetRatio(meta *meta.Meta) *ratio.Ratio {
+ return adaptor.GetRatioHelper(meta, RatioMap)
+}
+
func (a *Adaptor) GetModelList() []string {
- return ModelList
+ return adaptor.GetModelListHelper(RatioMap)
}
func (a *Adaptor) GetChannelName() string {
diff --git a/relay/adaptor/coze/constants.go b/relay/adaptor/coze/constants.go
index d20fd875..e98da846 100644
--- a/relay/adaptor/coze/constants.go
+++ b/relay/adaptor/coze/constants.go
@@ -1,3 +1,5 @@
package coze
-var ModelList = []string{}
+import "github.com/songquanpeng/one-api/relay/billing/ratio"
+
+var RatioMap = map[string]ratio.Ratio{}
diff --git a/relay/adaptor/deepl/adaptor.go b/relay/adaptor/deepl/adaptor.go
index d018a096..8e413672 100644
--- a/relay/adaptor/deepl/adaptor.go
+++ b/relay/adaptor/deepl/adaptor.go
@@ -3,12 +3,14 @@ package deepl
import (
"errors"
"fmt"
- "github.com/gin-gonic/gin"
- "github.com/songquanpeng/one-api/relay/adaptor"
- "github.com/songquanpeng/one-api/relay/meta"
- "github.com/songquanpeng/one-api/relay/model"
"io"
"net/http"
+
+ "github.com/gin-gonic/gin"
+ "github.com/songquanpeng/one-api/relay/adaptor"
+ "github.com/songquanpeng/one-api/relay/billing/ratio"
+ "github.com/songquanpeng/one-api/relay/meta"
+ "github.com/songquanpeng/one-api/relay/model"
)
type Adaptor struct {
@@ -64,8 +66,12 @@ func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Met
return
}
+func (a *Adaptor) GetRatio(meta *meta.Meta) *ratio.Ratio {
+ return adaptor.GetRatioHelper(meta, RatioMap)
+}
+
func (a *Adaptor) GetModelList() []string {
- return ModelList
+ return adaptor.GetModelListHelper(RatioMap)
}
func (a *Adaptor) GetChannelName() string {
diff --git a/relay/adaptor/deepl/constants.go b/relay/adaptor/deepl/constants.go
index 6a4f2545..f98b8c84 100644
--- a/relay/adaptor/deepl/constants.go
+++ b/relay/adaptor/deepl/constants.go
@@ -1,9 +1,10 @@
package deepl
-// https://developers.deepl.com/docs/api-reference/glossaries
+import "github.com/songquanpeng/one-api/relay/billing/ratio"
-var ModelList = []string{
- "deepl-zh",
- "deepl-en",
- "deepl-ja",
+// https://developers.deepl.com/docs/api-reference/glossaries
+var RatioMap = map[string]ratio.Ratio{
+ "deepl-zh": {Input: 25.0 * ratio.MILLI_USD, Output: 0},
+ "deepl-en": {Input: 25.0 * ratio.MILLI_USD, Output: 0},
+ "deepl-ja": {Input: 25.0 * ratio.MILLI_USD, Output: 0},
}
diff --git a/relay/adaptor/deepseek/constants.go b/relay/adaptor/deepseek/constants.go
index ad840bc2..57cc24ad 100644
--- a/relay/adaptor/deepseek/constants.go
+++ b/relay/adaptor/deepseek/constants.go
@@ -1,6 +1,8 @@
package deepseek
-var ModelList = []string{
- "deepseek-chat",
- "deepseek-coder",
+import "github.com/songquanpeng/one-api/relay/billing/ratio"
+
+var RatioMap = map[string]ratio.Ratio{
+ "deepseek-chat": {Input: 1 * ratio.MILLI_RMB, Output: 2 * ratio.MILLI_RMB},
+ "deepseek-coder": {Input: 1 * ratio.MILLI_RMB, Output: 2 * ratio.MILLI_RMB},
}
diff --git a/relay/adaptor/doubao/constants.go b/relay/adaptor/doubao/constants.go
index dbe819dd..78b81f86 100644
--- a/relay/adaptor/doubao/constants.go
+++ b/relay/adaptor/doubao/constants.go
@@ -1,13 +1,15 @@
package doubao
-// https://console.volcengine.com/ark/region:ark+cn-beijing/model
+import "github.com/songquanpeng/one-api/relay/billing/ratio"
-var ModelList = []string{
- "Doubao-pro-128k",
- "Doubao-pro-32k",
- "Doubao-pro-4k",
- "Doubao-lite-128k",
- "Doubao-lite-32k",
- "Doubao-lite-4k",
- "Doubao-embedding",
+// https://www.volcengine.com/product/doubao
+var RatioMap = map[string]ratio.Ratio{
+ "Doubao-vision-pro-32k": {Input: 0.0030 * ratio.RMB, Output: 0.0090 * ratio.RMB},
+ "Doubao-vision-lite-32k": {Input: 0.0015 * ratio.RMB, Output: 0.0045 * ratio.RMB},
+ "Doubao-pro-256k": {Input: 0.0050 * ratio.RMB, Output: 0.0090 * ratio.RMB},
+ "Doubao-pro-128k": {Input: 0.0050 * ratio.RMB, Output: 0.0090 * ratio.RMB},
+ "Doubao-pro-32k": {Input: 0.0008 * ratio.RMB, Output: 0.0020 * ratio.RMB},
+ "Doubao-lite-128k": {Input: 0.0008 * ratio.RMB, Output: 0.0010 * ratio.RMB},
+ "Doubao-lite-32k": {Input: 0.0003 * ratio.RMB, Output: 0.0006 * ratio.RMB},
+ "Doubao-embedding": {Input: 0.0005 * ratio.RMB, Output: 0},
}
diff --git a/relay/adaptor/gemini/adaptor.go b/relay/adaptor/gemini/adaptor.go
index a86fde40..dbe70a03 100644
--- a/relay/adaptor/gemini/adaptor.go
+++ b/relay/adaptor/gemini/adaptor.go
@@ -11,6 +11,7 @@ import (
"github.com/songquanpeng/one-api/common/helper"
channelhelper "github.com/songquanpeng/one-api/relay/adaptor"
"github.com/songquanpeng/one-api/relay/adaptor/openai"
+ "github.com/songquanpeng/one-api/relay/billing/ratio"
"github.com/songquanpeng/one-api/relay/meta"
"github.com/songquanpeng/one-api/relay/model"
"github.com/songquanpeng/one-api/relay/relaymode"
@@ -92,8 +93,12 @@ func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Met
return
}
+func (a *Adaptor) GetRatio(meta *meta.Meta) *ratio.Ratio {
+ return channelhelper.GetRatioHelper(meta, RatioMap)
+}
+
func (a *Adaptor) GetModelList() []string {
- return ModelList
+ return channelhelper.GetModelListHelper(RatioMap)
}
func (a *Adaptor) GetChannelName() string {
diff --git a/relay/adaptor/gemini/constants.go b/relay/adaptor/gemini/constants.go
index 9d1cbc4a..6e429ada 100644
--- a/relay/adaptor/gemini/constants.go
+++ b/relay/adaptor/gemini/constants.go
@@ -1,11 +1,53 @@
package gemini
+import "github.com/songquanpeng/one-api/relay/billing/ratio"
+
// 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 gemini15FlashRatio = ratio.Ratio{
+ Input: 0.075 * ratio.MILLI_USD,
+ Output: 0.30 * ratio.MILLI_USD,
+ LongThreshold: 128000,
+ LongInput: 0.15 * ratio.MILLI_USD,
+ LongOutput: 0.60 * ratio.MILLI_USD,
+}
+
+var gemini15ProRatio = ratio.Ratio{
+ Input: 1.25 * ratio.MILLI_USD,
+ Output: 5.00 * ratio.MILLI_USD,
+ LongThreshold: 128000,
+ LongInput: 2.50 * ratio.MILLI_USD,
+ LongOutput: 10.00 * ratio.MILLI_USD,
+}
+
+var gemini10ProRatio = ratio.Ratio{
+ Input: 0.50 * ratio.MILLI_USD,
+ Output: 1.50 * ratio.MILLI_USD,
+}
+
+var gemini15Flash8bRatio = ratio.Ratio{
+ Input: 0.0375 * ratio.MILLI_USD,
+ Output: 0.15 * ratio.MILLI_USD,
+ LongThreshold: 128000,
+ LongInput: 0.075 * ratio.MILLI_USD,
+ LongOutput: 0.30 * ratio.MILLI_USD,
+}
+
+// https://ai.google.dev/pricing
+// https://ai.google.dev/gemini-api/docs/models/gemini
+// https://cloud.google.com/vertex-ai/generative-ai/pricing?hl=zh-cn#google_models
+var RatioMap = map[string]ratio.Ratio{
+ "gemini-2.0-flash-exp": {Input: 0.1, Output: 0.1}, // currently free of charge
+ "gemini-2.0-flash-thinking-exp": {Input: 0.1, Output: 0.1}, // currently free of charge
+ "gemini-1.5-flash": gemini15FlashRatio,
+ "gemini-1.5-flash-001": gemini15FlashRatio,
+ "gemini-1.5-flash-002": gemini15FlashRatio,
+ "gemini-1.5-pro": gemini15ProRatio,
+ "gemini-1.5-pro-001": gemini15ProRatio,
+ "gemini-1.5-pro-002": gemini15ProRatio,
+ "gemini-1.0-pro": gemini10ProRatio,
+ "gemini-1.0-pro-001": gemini10ProRatio,
+ "gemini-1.5-flash-8b": gemini15Flash8bRatio,
+ "gemini-1.5-flash-8b-001": gemini15Flash8bRatio,
+ "text-embedding-004": {Input: 0.1, Output: 0.1}, // free of charge
}
diff --git a/relay/adaptor/groq/constants.go b/relay/adaptor/groq/constants.go
index 0864ebe7..7e395b7a 100644
--- a/relay/adaptor/groq/constants.go
+++ b/relay/adaptor/groq/constants.go
@@ -1,27 +1,18 @@
package groq
-// https://console.groq.com/docs/models
+import "github.com/songquanpeng/one-api/relay/billing/ratio"
-var ModelList = []string{
- "gemma-7b-it",
- "gemma2-9b-it",
- "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",
+// https://groq.com/pricing/
+// https://console.groq.com/docs/models
+var RatioMap = map[string]ratio.Ratio{
+ "distil-whisper-large-v3-en": {Input: 0.02 / 3600 * 20 * ratio.USD, Output: 0},
+ "gemma2-9b-it": {Input: 0.20 * ratio.MILLI_USD, Output: 0.20 * ratio.MILLI_USD},
+ "llama-3.3-70b-versatile": {Input: 0.59 * ratio.MILLI_USD, Output: 0.79 * ratio.MILLI_USD},
+ "llama-3.1-8b-instant": {Input: 0.05 * ratio.MILLI_USD, Output: 0.08 * ratio.MILLI_USD},
+ "llama-guard-3-8b": {Input: 0.20 * ratio.MILLI_USD, Output: 0.20 * ratio.MILLI_USD},
+ "llama3-70b-8192": {Input: 0.59 * ratio.MILLI_USD, Output: 0.79 * ratio.MILLI_USD},
+ "llama3-8b-8192": {Input: 0.05 * ratio.MILLI_USD, Output: 0.08 * ratio.MILLI_USD},
+ "mixtral-8x7b-32768": {Input: 0.24 * ratio.MILLI_USD, Output: 0.24 * ratio.MILLI_USD},
+ "whisper-large-v3": {Input: 0.111 / 3600 * 20 * ratio.USD, Output: 0},
+ "whisper-large-v3-turbo": {Input: 0.04 / 3600 * 20 * ratio.USD, Output: 0},
}
diff --git a/relay/adaptor/interface.go b/relay/adaptor/interface.go
index 01b2e2cb..cdd89bf9 100644
--- a/relay/adaptor/interface.go
+++ b/relay/adaptor/interface.go
@@ -2,6 +2,7 @@ package adaptor
import (
"github.com/gin-gonic/gin"
+ "github.com/songquanpeng/one-api/relay/billing/ratio"
"github.com/songquanpeng/one-api/relay/meta"
"github.com/songquanpeng/one-api/relay/model"
"io"
@@ -16,6 +17,7 @@ type Adaptor interface {
ConvertImageRequest(request *model.ImageRequest) (any, error)
DoRequest(c *gin.Context, meta *meta.Meta, requestBody io.Reader) (*http.Response, error)
DoResponse(c *gin.Context, resp *http.Response, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode)
+ GetRatio(meta *meta.Meta) *ratio.Ratio
GetModelList() []string
GetChannelName() string
}
diff --git a/relay/adaptor/lingyiwanwu/constants.go b/relay/adaptor/lingyiwanwu/constants.go
index 30000e9d..6185614f 100644
--- a/relay/adaptor/lingyiwanwu/constants.go
+++ b/relay/adaptor/lingyiwanwu/constants.go
@@ -1,9 +1,9 @@
package lingyiwanwu
-// https://platform.lingyiwanwu.com/docs
+import "github.com/songquanpeng/one-api/relay/billing/ratio"
-var ModelList = []string{
- "yi-34b-chat-0205",
- "yi-34b-chat-200k",
- "yi-vl-plus",
+// https://platform.lingyiwanwu.com/docs#%E6%A8%A1%E5%9E%8B%E4%B8%8E%E8%AE%A1%E8%B4%B9
+var RatioMap = map[string]ratio.Ratio{
+ "yi-lightning": {Input: 0.99 * ratio.MILLI_RMB, Output: 0.99 * ratio.MILLI_RMB},
+ "yi-vision-v2": {Input: 6 * ratio.MILLI_RMB, Output: 6 * ratio.MILLI_RMB},
}
diff --git a/relay/adaptor/minimax/constants.go b/relay/adaptor/minimax/constants.go
index 1b2fc104..0e780e1e 100644
--- a/relay/adaptor/minimax/constants.go
+++ b/relay/adaptor/minimax/constants.go
@@ -1,11 +1,15 @@
package minimax
-// https://www.minimaxi.com/document/guides/chat-model/V2?id=65e0736ab2845de20908e2dd
+import "github.com/songquanpeng/one-api/relay/billing/ratio"
-var ModelList = []string{
- "abab6.5-chat",
- "abab6.5s-chat",
- "abab6-chat",
- "abab5.5-chat",
- "abab5.5s-chat",
+// https://platform.minimaxi.com/document/Price
+// https://platform.minimaxi.com/document/ChatCompletion%20v2
+
+var RatioMap = map[string]ratio.Ratio{
+ "abab7-chat-preview": {Input: 0.01 * ratio.RMB, Output: 0.01 * ratio.RMB},
+ "abab6.5s-chat": {Input: 0.001 * ratio.RMB, Output: 0.001 * ratio.RMB},
+ "abab6.5g-chat": {Input: 0.005 * ratio.RMB, Output: 0.005 * ratio.RMB},
+ "abab6.5t-chat": {Input: 0.005 * ratio.RMB, Output: 0.005 * ratio.RMB},
+ "abab5.5s-chat": {Input: 0.005 * ratio.RMB, Output: 0.005 * ratio.RMB},
+ "abab5.5-chat": {Input: 0.015 * ratio.RMB, Output: 0.015 * ratio.RMB},
}
diff --git a/relay/adaptor/mistral/constants.go b/relay/adaptor/mistral/constants.go
index cdb157f5..2d243267 100644
--- a/relay/adaptor/mistral/constants.go
+++ b/relay/adaptor/mistral/constants.go
@@ -1,10 +1,20 @@
package mistral
-var ModelList = []string{
- "open-mistral-7b",
- "open-mixtral-8x7b",
- "mistral-small-latest",
- "mistral-medium-latest",
- "mistral-large-latest",
- "mistral-embed",
+import "github.com/songquanpeng/one-api/relay/billing/ratio"
+
+// https://mistral.ai/technology/#pricing
+var RatioMap = map[string]ratio.Ratio{
+ "mistral-large-latest": {Input: 2 * ratio.MILLI_USD, Output: 6 * ratio.MILLI_USD},
+ "pixtral-large-latest": {Input: 2 * ratio.MILLI_USD, Output: 6 * ratio.MILLI_USD},
+ "mistral-small-latest": {Input: 0.2 * ratio.MILLI_USD, Output: 0.6 * ratio.MILLI_USD},
+ "codestral-latest": {Input: 0.2 * ratio.MILLI_USD, Output: 0.6 * ratio.MILLI_USD},
+ "ministral-8b-latest": {Input: 0.1 * ratio.MILLI_USD, Output: 0.1 * ratio.MILLI_USD},
+ "ministral-3b-latest": {Input: 0.04 * ratio.MILLI_USD, Output: 0.04 * ratio.MILLI_USD},
+ "mistral-embed": {Input: 0.1 * ratio.MILLI_USD, Output: 0},
+ "mistral-moderation-latest": {Input: 0.1 * ratio.MILLI_USD, Output: 0},
+ "pixtral-12b": {Input: 0.15 * ratio.MILLI_USD, Output: 0.15 * ratio.MILLI_USD},
+ "mistral-nemo": {Input: 0.15 * ratio.MILLI_USD, Output: 0.15 * ratio.MILLI_USD},
+ "open-mistral-7b": {Input: 0.25 * ratio.MILLI_USD, Output: 0.25 * ratio.MILLI_USD},
+ "open-mixtral-8x7b": {Input: 0.7 * ratio.MILLI_USD, Output: 0.7 * ratio.MILLI_USD},
+ "open-mixtral-8x22b": {Input: 2 * ratio.MILLI_USD, Output: 6 * ratio.MILLI_USD},
}
diff --git a/relay/adaptor/moonshot/constants.go b/relay/adaptor/moonshot/constants.go
index 1b86f0fa..6f97e1ca 100644
--- a/relay/adaptor/moonshot/constants.go
+++ b/relay/adaptor/moonshot/constants.go
@@ -1,7 +1,10 @@
package moonshot
-var ModelList = []string{
- "moonshot-v1-8k",
- "moonshot-v1-32k",
- "moonshot-v1-128k",
+import "github.com/songquanpeng/one-api/relay/billing/ratio"
+
+// https://platform.moonshot.cn/docs/pricing/chat#%E4%BA%A7%E5%93%81%E5%AE%9A%E4%BB%B7
+var RatioMap = map[string]ratio.Ratio{
+ "moonshot-v1-8k": {Input: 12 * ratio.MILLI_RMB, Output: 12 * ratio.MILLI_RMB},
+ "moonshot-v1-32k": {Input: 24 * ratio.MILLI_RMB, Output: 24 * ratio.MILLI_RMB},
+ "moonshot-v1-128k": {Input: 60 * ratio.MILLI_RMB, Output: 60 * ratio.MILLI_RMB},
}
diff --git a/relay/adaptor/novita/constants.go b/relay/adaptor/novita/constants.go
index c6618308..cf5c6210 100644
--- a/relay/adaptor/novita/constants.go
+++ b/relay/adaptor/novita/constants.go
@@ -1,19 +1,38 @@
package novita
+import "github.com/songquanpeng/one-api/relay/billing/ratio"
+
// https://novita.ai/llm-api
-var ModelList = []string{
- "meta-llama/llama-3-8b-instruct",
- "meta-llama/llama-3-70b-instruct",
- "nousresearch/hermes-2-pro-llama-3-8b",
- "nousresearch/nous-hermes-llama2-13b",
- "mistralai/mistral-7b-instruct",
- "cognitivecomputations/dolphin-mixtral-8x22b",
- "sao10k/l3-70b-euryale-v2.1",
- "sophosympatheia/midnight-rose-70b",
- "gryphe/mythomax-l2-13b",
- "Nous-Hermes-2-Mixtral-8x7B-DPO",
- "lzlv_70b",
- "teknium/openhermes-2.5-mistral-7b",
- "microsoft/wizardlm-2-8x22b",
+var RatioMap = map[string]ratio.Ratio{
+ "meta-llama/llama-3.3-70b-instruct": {Input: 0.39 * ratio.MILLI_USD, Output: 0.39 * ratio.MILLI_USD},
+ "meta-llama/llama-3.1-8b-instruct": {Input: 0.05 * ratio.MILLI_USD, Output: 0.05 * ratio.MILLI_USD},
+ "meta-llama/llama-3.1-8b-instruct-max": {Input: 0.05 * ratio.MILLI_USD, Output: 0.05 * ratio.MILLI_USD},
+ "meta-llama/llama-3.1-70b-instruct": {Input: 0.34 * ratio.MILLI_USD, Output: 0.39 * ratio.MILLI_USD},
+ "meta-llama/llama-3-8b-instruct": {Input: 0.04 * ratio.MILLI_USD, Output: 0.04 * ratio.MILLI_USD},
+ "meta-llama/llama-3-70b-instruct": {Input: 0.51 * ratio.MILLI_USD, Output: 0.74 * ratio.MILLI_USD},
+ "gryphe/mythomax-l2-13b": {Input: 0.09 * ratio.MILLI_USD, Output: 0.09 * ratio.MILLI_USD},
+ "google/gemma-2-9b-it": {Input: 0.08 * ratio.MILLI_USD, Output: 0.08 * ratio.MILLI_USD},
+ "mistralai/mistral-nemo": {Input: 0.17 * ratio.MILLI_USD, Output: 0.17 * ratio.MILLI_USD},
+ "microsoft/wizardlm-2-8x22b": {Input: 0.62 * ratio.MILLI_USD, Output: 0.62 * ratio.MILLI_USD},
+ "mistralai/mistral-7b-instruct": {Input: 0.059 * ratio.MILLI_USD, Output: 0.059 * ratio.MILLI_USD},
+ "openchat/openchat-7b": {Input: 0.06 * ratio.MILLI_USD, Output: 0.06 * ratio.MILLI_USD},
+ "nousresearch/hermes-2-pro-llama-3-8b": {Input: 0.14 * ratio.MILLI_USD, Output: 0.14 * ratio.MILLI_USD},
+ "sao10k/l3-70b-euryale-v2.1": {Input: 1.48 * ratio.MILLI_USD, Output: 1.48 * ratio.MILLI_USD},
+ "cognitivecomputations/dolphin-mixtral-8x22b": {Input: 0.9 * ratio.MILLI_USD, Output: 0.9 * ratio.MILLI_USD},
+ "jondurbin/airoboros-l2-70b": {Input: 0.5 * ratio.MILLI_USD, Output: 0.5 * ratio.MILLI_USD},
+ "nousresearch/nous-hermes-llama2-13b": {Input: 0.17 * ratio.MILLI_USD, Output: 0.17 * ratio.MILLI_USD},
+ "teknium/openhermes-2.5-mistral-7b": {Input: 0.17 * ratio.MILLI_USD, Output: 0.17 * ratio.MILLI_USD},
+ "sophosympatheia/midnight-rose-70b": {Input: 0.8 * ratio.MILLI_USD, Output: 0.8 * ratio.MILLI_USD},
+ "Sao10K/L3-8B-Stheno-v3.2": {Input: 0.05 * ratio.MILLI_USD, Output: 0.05 * ratio.MILLI_USD},
+ "sao10k/l3-8b-lunaris": {Input: 0.05 * ratio.MILLI_USD, Output: 0.05 * ratio.MILLI_USD},
+ "qwen/qwen-2-vl-72b-instruct": {Input: 0.45 * ratio.MILLI_USD, Output: 0.45 * ratio.MILLI_USD},
+ "meta-llama/llama-3.2-1b-instruct": {Input: 0.02 * ratio.MILLI_USD, Output: 0.02 * ratio.MILLI_USD},
+ "meta-llama/llama-3.2-11b-vision-instruct": {Input: 0.06 * ratio.MILLI_USD, Output: 0.06 * ratio.MILLI_USD},
+ "meta-llama/llama-3.2-3b-instruct": {Input: 0.03 * ratio.MILLI_USD, Output: 0.05 * ratio.MILLI_USD},
+ "meta-llama/llama-3.1-8b-instruct-bf16": {Input: 0.06 * ratio.MILLI_USD, Output: 0.06 * ratio.MILLI_USD},
+ "qwen/qwen-2.5-72b-instruct": {Input: 0.38 * ratio.MILLI_USD, Output: 0.4 * ratio.MILLI_USD},
+ "sao10k/l31-70b-euryale-v2.2": {Input: 1.48 * ratio.MILLI_USD, Output: 1.48 * ratio.MILLI_USD},
+ "qwen/qwen-2-7b-instruct": {Input: 0.054 * ratio.MILLI_USD, Output: 0.054 * ratio.MILLI_USD},
+ "qwen/qwen-2-72b-instruct": {Input: 0.34 * ratio.MILLI_USD, Output: 0.39 * ratio.MILLI_USD},
}
diff --git a/relay/adaptor/ollama/adaptor.go b/relay/adaptor/ollama/adaptor.go
index ad1f8983..cad86216 100644
--- a/relay/adaptor/ollama/adaptor.go
+++ b/relay/adaptor/ollama/adaptor.go
@@ -3,11 +3,13 @@ package ollama
import (
"errors"
"fmt"
- "github.com/songquanpeng/one-api/relay/meta"
- "github.com/songquanpeng/one-api/relay/relaymode"
"io"
"net/http"
+ "github.com/songquanpeng/one-api/relay/billing/ratio"
+ "github.com/songquanpeng/one-api/relay/meta"
+ "github.com/songquanpeng/one-api/relay/relaymode"
+
"github.com/gin-gonic/gin"
"github.com/songquanpeng/one-api/relay/adaptor"
"github.com/songquanpeng/one-api/relay/model"
@@ -73,8 +75,12 @@ func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Met
return
}
+func (a *Adaptor) GetRatio(meta *meta.Meta) *ratio.Ratio {
+ return adaptor.GetRatioHelper(meta, RatioMap)
+}
+
func (a *Adaptor) GetModelList() []string {
- return ModelList
+ return adaptor.GetModelListHelper(RatioMap)
}
func (a *Adaptor) GetChannelName() string {
diff --git a/relay/adaptor/ollama/constants.go b/relay/adaptor/ollama/constants.go
index d9dc72a8..97ee4f47 100644
--- a/relay/adaptor/ollama/constants.go
+++ b/relay/adaptor/ollama/constants.go
@@ -1,11 +1,13 @@
package ollama
-var ModelList = []string{
- "codellama:7b-instruct",
- "llama2:7b",
- "llama2:latest",
- "llama3:latest",
- "phi3:latest",
- "qwen:0.5b-chat",
- "qwen:7b",
+import "github.com/songquanpeng/one-api/relay/billing/ratio"
+
+var RatioMap = map[string]ratio.Ratio{
+ "codellama:7b-instruct": {Input: 0, Output: 0},
+ "llama2:7b": {Input: 0, Output: 0},
+ "llama2:latest": {Input: 0, Output: 0},
+ "llama3:latest": {Input: 0, Output: 0},
+ "phi3:latest": {Input: 0, Output: 0},
+ "qwen:0.5b-chat": {Input: 0, Output: 0},
+ "qwen:7b": {Input: 0, Output: 0},
}
diff --git a/relay/adaptor/openai/adaptor.go b/relay/adaptor/openai/adaptor.go
index 6946e402..88b2579d 100644
--- a/relay/adaptor/openai/adaptor.go
+++ b/relay/adaptor/openai/adaptor.go
@@ -12,6 +12,7 @@ import (
"github.com/songquanpeng/one-api/relay/adaptor/doubao"
"github.com/songquanpeng/one-api/relay/adaptor/minimax"
"github.com/songquanpeng/one-api/relay/adaptor/novita"
+ "github.com/songquanpeng/one-api/relay/billing/ratio"
"github.com/songquanpeng/one-api/relay/channeltype"
"github.com/songquanpeng/one-api/relay/meta"
"github.com/songquanpeng/one-api/relay/model"
@@ -118,9 +119,14 @@ func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Met
return
}
+func (a *Adaptor) GetRatio(meta *meta.Meta) *ratio.Ratio {
+ _, ratioMap := GetCompatibleChannelMeta(a.ChannelType)
+ return adaptor.GetRatioHelper(meta, ratioMap)
+}
+
func (a *Adaptor) GetModelList() []string {
- _, modelList := GetCompatibleChannelMeta(a.ChannelType)
- return modelList
+ _, ratioMap := GetCompatibleChannelMeta(a.ChannelType)
+ return adaptor.GetModelListHelper(ratioMap)
}
func (a *Adaptor) GetChannelName() string {
diff --git a/relay/adaptor/openai/compatible.go b/relay/adaptor/openai/compatible.go
index 15b4dcc0..050b97d2 100644
--- a/relay/adaptor/openai/compatible.go
+++ b/relay/adaptor/openai/compatible.go
@@ -15,6 +15,7 @@ import (
"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/billing/ratio"
"github.com/songquanpeng/one-api/relay/channeltype"
)
@@ -36,39 +37,39 @@ var CompatibleChannels = []int{
channeltype.XAI,
}
-func GetCompatibleChannelMeta(channelType int) (string, []string) {
+func GetCompatibleChannelMeta(channelType int) (string, map[string]ratio.Ratio) {
switch channelType {
case channeltype.Azure:
- return "azure", ModelList
+ return "azure", RatioMap
case channeltype.AI360:
- return "360", ai360.ModelList
+ return "360", ai360.RatioMap
case channeltype.Moonshot:
- return "moonshot", moonshot.ModelList
+ return "moonshot", moonshot.RatioMap
case channeltype.Baichuan:
- return "baichuan", baichuan.ModelList
+ return "baichuan", baichuan.RatioMap
case channeltype.Minimax:
- return "minimax", minimax.ModelList
+ return "minimax", minimax.RatioMap
case channeltype.Mistral:
- return "mistralai", mistral.ModelList
+ return "mistralai", mistral.RatioMap
case channeltype.Groq:
- return "groq", groq.ModelList
+ return "groq", groq.RatioMap
case channeltype.LingYiWanWu:
- return "lingyiwanwu", lingyiwanwu.ModelList
+ return "lingyiwanwu", lingyiwanwu.RatioMap
case channeltype.StepFun:
- return "stepfun", stepfun.ModelList
+ return "stepfun", stepfun.RatioMap
case channeltype.DeepSeek:
- return "deepseek", deepseek.ModelList
+ return "deepseek", deepseek.RatioMap
case channeltype.TogetherAI:
- return "together.ai", togetherai.ModelList
+ return "together.ai", togetherai.RatioMap
case channeltype.Doubao:
- return "doubao", doubao.ModelList
+ return "doubao", doubao.RatioMap
case channeltype.Novita:
- return "novita", novita.ModelList
+ return "novita", novita.RatioMap
case channeltype.SiliconFlow:
- return "siliconflow", siliconflow.ModelList
+ return "siliconflow", siliconflow.RatioMap
case channeltype.XAI:
- return "xai", xai.ModelList
+ return "xai", xai.RatioMap
default:
- return "openai", ModelList
+ return "openai", RatioMap
}
}
diff --git a/relay/adaptor/openai/constants.go b/relay/adaptor/openai/constants.go
index 8a643bc6..f749e1e6 100644
--- a/relay/adaptor/openai/constants.go
+++ b/relay/adaptor/openai/constants.go
@@ -1,27 +1,61 @@
package openai
-var ModelList = []string{
- "gpt-3.5-turbo", "gpt-3.5-turbo-0301", "gpt-3.5-turbo-0613", "gpt-3.5-turbo-1106", "gpt-3.5-turbo-0125",
- "gpt-3.5-turbo-16k", "gpt-3.5-turbo-16k-0613",
- "gpt-3.5-turbo-instruct",
- "gpt-4", "gpt-4-0314", "gpt-4-0613", "gpt-4-1106-preview", "gpt-4-0125-preview",
- "gpt-4-32k", "gpt-4-32k-0314", "gpt-4-32k-0613",
- "gpt-4-turbo-preview", "gpt-4-turbo", "gpt-4-turbo-2024-04-09",
- "gpt-4o", "gpt-4o-2024-05-13",
- "gpt-4o-2024-08-06",
- "gpt-4o-2024-11-20",
- "chatgpt-4o-latest",
- "gpt-4o-mini", "gpt-4o-mini-2024-07-18",
- "gpt-4-vision-preview",
- "text-embedding-ada-002", "text-embedding-3-small", "text-embedding-3-large",
- "text-curie-001", "text-babbage-001", "text-ada-001", "text-davinci-002", "text-davinci-003",
- "text-moderation-latest", "text-moderation-stable",
- "text-davinci-edit-001",
- "davinci-002", "babbage-002",
- "dall-e-2", "dall-e-3",
- "whisper-1",
- "tts-1", "tts-1-1106", "tts-1-hd", "tts-1-hd-1106",
- "o1", "o1-2024-12-17",
- "o1-preview", "o1-preview-2024-09-12",
- "o1-mini", "o1-mini-2024-09-12",
+import "github.com/songquanpeng/one-api/relay/billing/ratio"
+
+var RatioMap = map[string]ratio.Ratio{
+ "gpt-4": {Input: 15, Output: 30},
+ "gpt-4-0314": {Input: 15, Output: 30},
+ "gpt-4-0613": {Input: 15, Output: 30},
+ "gpt-4-32k": {Input: 30, Output: 60},
+ "gpt-4-32k-0314": {Input: 30, Output: 60},
+ "gpt-4-32k-0613": {Input: 30, Output: 60},
+ "gpt-4-1106-preview": {Input: 5, Output: 15},
+ "gpt-4-0125-preview": {Input: 5, Output: 15},
+ "gpt-4-turbo-preview": {Input: 5, Output: 15}, // $0.01 / 1K tokens
+ "gpt-4-turbo": {Input: 5, Output: 15}, // $0.01 / 1K tokens
+ "gpt-4-turbo-2024-04-09": {Input: 5, Output: 15}, // $0.01 / 1K tokens
+ "gpt-4o": {Input: 1.25, Output: 5}, // $0.005 / 1K tokens
+ "chatgpt-4o-latest": {Input: 2.5, Output: 7.5}, // $0.005 / 1K tokens
+ "gpt-4o-2024-05-13": {Input: 2.5, Output: 7.5}, // $0.005 / 1K tokens
+ "gpt-4o-2024-08-06": {Input: 1.25, Output: 5}, // $0.0025 / 1K tokens
+ "gpt-4o-2024-11-20": {Input: 1.25, Output: 5}, // $0.0025 / 1K tokens
+ "gpt-4o-mini": {Input: 0.075, Output: 0.3}, // $0.00015 / 1K tokens
+ "gpt-4o-mini-2024-07-18": {Input: 0.075, Output: 0.3}, // $0.00015 / 1K tokens
+ "gpt-4-vision-preview": {Input: 5, Output: 15}, // $0.01 / 1K tokens
+ "gpt-3.5-turbo": {Input: 0.25, Output: 0.75}, // $0.0005 / 1K tokens
+ "gpt-3.5-turbo-0301": {Input: 0.75, Output: 1}, // $0.0015 / 1K tokens
+ "gpt-3.5-turbo-0613": {Input: 0.75, Output: 1}, // $0.0015 / 1K tokens
+ "gpt-3.5-turbo-16k": {Input: 1.5, Output: 2}, // $0.003 / 1K tokens
+ "gpt-3.5-turbo-16k-0613": {Input: 1.5, Output: 2}, // $0.003 / 1K tokens
+ "gpt-3.5-turbo-instruct": {Input: 0.75, Output: 1}, // $0.0015 / 1K tokens
+ "gpt-3.5-turbo-1106": {Input: 0.5, Output: 1}, // $0.001 / 1K tokens
+ "gpt-3.5-turbo-0125": {Input: 0.25, Output: 0.75}, // $0.0005 / 1K tokens
+ "davinci-002": {Input: 1, Output: 1}, // $0.002 / 1K tokens
+ "babbage-002": {Input: 0.2, Output: 0.2}, // $0.0004 / 1K tokens
+ "text-ada-001": {Input: 0.2, Output: 0.2}, // $0.0004 / 1K tokens
+ "text-babbage-001": {Input: 0.25, Output: 0.25}, // $0.0005 / 1K tokens
+ "text-curie-001": {Input: 1, Output: 1}, // $0.002 / 1K tokens
+ "text-davinci-002": {Input: 10, Output: 10}, // $0.02 / 1K tokens
+ "text-davinci-003": {Input: 10, Output: 10}, // $0.02 / 1K tokens
+ "text-davinci-edit-001": {Input: 10, Output: 10}, // $0.02 / 1K tokens
+ "code-davinci-edit-001": {Input: 10, Output: 10}, // $0.02 / 1K tokens
+ "whisper-1": {Input: 1, Output: 1}, // $0.006 / minute -> $0.002 / 20 seconds -> $0.002 / 1K tokens -> 20 seconds / 1K tokens
+ "tts-1": {Input: 7.5, Output: 7.5}, // $0.015 / 1K characters
+ "tts-1-1106": {Input: 7.5, Output: 7.5}, // $0.015 / 1K characters
+ "tts-1-hd": {Input: 15, Output: 15}, // $0.030 / 1K characters
+ "tts-1-hd-1106": {Input: 15, Output: 15}, // $0.030 / 1K characters
+ "davinci": {Input: 10, Output: 10}, // $0.02 / 1K tokens
+ "curie": {Input: 10, Output: 10}, // $0.02 / 1K tokens
+ "babbage": {Input: 10, Output: 10}, // $0.02 / 1K tokens
+ "ada": {Input: 10, Output: 10}, // $0.02 / 1K tokens
+ "text-embedding-ada-002": {Input: 0.05, Output: 0}, // $0.001 / 1K tokens
+ "text-embedding-3-small": {Input: 0.01, Output: 0}, // $0.0002 / 1K tokens
+ "text-embedding-3-large": {Input: 0.065, Output: 0}, // $0.0013 / 1K tokens
+ "text-search-ada-doc-001": {Input: 10, Output: 0}, // $0.02 / 1K tokens
+ "text-moderation-stable": {Input: 0.1, Output: 0}, // $0.002 / 1K tokens
+ "text-moderation-latest": {Input: 0.1, Output: 0}, // $0.002 / 1K tokens
+ "omni-moderation-latest": {Input: 0.1, Output: 0}, // $0.002 / 1K tokens
+ "omni-moderation-2024-09-26": {Input: 0.1, Output: 0}, // $0.002 / 1K tokens
+ "dall-e-2": {Input: 0.02 * ratio.USD, Output: 0}, // $0.016 - $0.020 / image
+ "dall-e-3": {Input: 0.04 * ratio.USD, Output: 0}, // $0.040 - $0.120 / image
}
diff --git a/relay/adaptor/palm/adaptor.go b/relay/adaptor/palm/adaptor.go
index 98aa3e18..51a615d2 100644
--- a/relay/adaptor/palm/adaptor.go
+++ b/relay/adaptor/palm/adaptor.go
@@ -3,13 +3,15 @@ package palm
import (
"errors"
"fmt"
+ "io"
+ "net/http"
+
"github.com/gin-gonic/gin"
"github.com/songquanpeng/one-api/relay/adaptor"
"github.com/songquanpeng/one-api/relay/adaptor/openai"
+ "github.com/songquanpeng/one-api/relay/billing/ratio"
"github.com/songquanpeng/one-api/relay/meta"
"github.com/songquanpeng/one-api/relay/model"
- "io"
- "net/http"
)
type Adaptor struct {
@@ -58,8 +60,12 @@ func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Met
return
}
+func (a *Adaptor) GetRatio(meta *meta.Meta) *ratio.Ratio {
+ return adaptor.GetRatioHelper(meta, RatioMap)
+}
+
func (a *Adaptor) GetModelList() []string {
- return ModelList
+ return adaptor.GetModelListHelper(RatioMap)
}
func (a *Adaptor) GetChannelName() string {
diff --git a/relay/adaptor/palm/constants.go b/relay/adaptor/palm/constants.go
index a8349362..70421002 100644
--- a/relay/adaptor/palm/constants.go
+++ b/relay/adaptor/palm/constants.go
@@ -1,5 +1,7 @@
package palm
-var ModelList = []string{
- "PaLM-2",
+import "github.com/songquanpeng/one-api/relay/billing/ratio"
+
+var RatioMap = map[string]ratio.Ratio{
+ "PaLM-2": {Input: 0, Output: 0},
}
diff --git a/relay/adaptor/proxy/adaptor.go b/relay/adaptor/proxy/adaptor.go
index 670c7628..798060cf 100644
--- a/relay/adaptor/proxy/adaptor.go
+++ b/relay/adaptor/proxy/adaptor.go
@@ -10,6 +10,7 @@ import (
"github.com/pkg/errors"
"github.com/songquanpeng/one-api/relay/adaptor"
channelhelper "github.com/songquanpeng/one-api/relay/adaptor"
+ "github.com/songquanpeng/one-api/relay/billing/ratio"
"github.com/songquanpeng/one-api/relay/meta"
"github.com/songquanpeng/one-api/relay/model"
relaymodel "github.com/songquanpeng/one-api/relay/model"
@@ -48,6 +49,10 @@ func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Met
return nil, nil
}
+func (a *Adaptor) GetRatio(meta *meta.Meta) *ratio.Ratio {
+ return adaptor.GetRatioHelper(meta, nil)
+}
+
func (a *Adaptor) GetModelList() (models []string) {
return nil
}
diff --git a/relay/adaptor/replicate/adaptor.go b/relay/adaptor/replicate/adaptor.go
index a60a7de3..a54780a0 100644
--- a/relay/adaptor/replicate/adaptor.go
+++ b/relay/adaptor/replicate/adaptor.go
@@ -4,7 +4,6 @@ import (
"fmt"
"io"
"net/http"
- "slices"
"strings"
"time"
@@ -13,6 +12,7 @@ import (
"github.com/songquanpeng/one-api/common/logger"
"github.com/songquanpeng/one-api/relay/adaptor"
"github.com/songquanpeng/one-api/relay/adaptor/openai"
+ "github.com/songquanpeng/one-api/relay/billing/ratio"
"github.com/songquanpeng/one-api/relay/meta"
"github.com/songquanpeng/one-api/relay/model"
"github.com/songquanpeng/one-api/relay/relaymode"
@@ -96,7 +96,7 @@ func (a *Adaptor) Init(meta *meta.Meta) {
}
func (a *Adaptor) GetRequestURL(meta *meta.Meta) (string, error) {
- if !slices.Contains(ModelList, meta.OriginModelName) {
+ if _, ok := RatioMap[meta.OriginModelName]; !ok {
return "", errors.Errorf("model %s not supported", meta.OriginModelName)
}
@@ -127,8 +127,12 @@ func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Met
return
}
+func (a *Adaptor) GetRatio(meta *meta.Meta) *ratio.Ratio {
+ return adaptor.GetRatioHelper(meta, RatioMap)
+}
+
func (a *Adaptor) GetModelList() []string {
- return ModelList
+ return adaptor.GetModelListHelper(RatioMap)
}
func (a *Adaptor) GetChannelName() string {
diff --git a/relay/adaptor/replicate/constant.go b/relay/adaptor/replicate/constant.go
index 989142c9..fddeb3f6 100644
--- a/relay/adaptor/replicate/constant.go
+++ b/relay/adaptor/replicate/constant.go
@@ -1,58 +1,49 @@
package replicate
-// ModelList is a list of models that can be used with Replicate.
-//
+import "github.com/songquanpeng/one-api/relay/billing/ratio"
+
// https://replicate.com/pricing
-var ModelList = []string{
- // -------------------------------------
- // image model
- // -------------------------------------
- "black-forest-labs/flux-1.1-pro",
- "black-forest-labs/flux-1.1-pro-ultra",
- "black-forest-labs/flux-canny-dev",
- "black-forest-labs/flux-canny-pro",
- "black-forest-labs/flux-depth-dev",
- "black-forest-labs/flux-depth-pro",
- "black-forest-labs/flux-dev",
- "black-forest-labs/flux-dev-lora",
- "black-forest-labs/flux-fill-dev",
- "black-forest-labs/flux-fill-pro",
- "black-forest-labs/flux-pro",
- "black-forest-labs/flux-redux-dev",
- "black-forest-labs/flux-redux-schnell",
- "black-forest-labs/flux-schnell",
- "black-forest-labs/flux-schnell-lora",
- "ideogram-ai/ideogram-v2",
- "ideogram-ai/ideogram-v2-turbo",
- "recraft-ai/recraft-v3",
- "recraft-ai/recraft-v3-svg",
- "stability-ai/stable-diffusion-3",
- "stability-ai/stable-diffusion-3.5-large",
- "stability-ai/stable-diffusion-3.5-large-turbo",
- "stability-ai/stable-diffusion-3.5-medium",
- // -------------------------------------
- // language model
- // -------------------------------------
- "ibm-granite/granite-20b-code-instruct-8k",
- "ibm-granite/granite-3.0-2b-instruct",
- "ibm-granite/granite-3.0-8b-instruct",
- "ibm-granite/granite-8b-code-instruct-128k",
- "meta/llama-2-13b",
- "meta/llama-2-13b-chat",
- "meta/llama-2-70b",
- "meta/llama-2-70b-chat",
- "meta/llama-2-7b",
- "meta/llama-2-7b-chat",
- "meta/meta-llama-3.1-405b-instruct",
- "meta/meta-llama-3-70b",
- "meta/meta-llama-3-70b-instruct",
- "meta/meta-llama-3-8b",
- "meta/meta-llama-3-8b-instruct",
- "mistralai/mistral-7b-instruct-v0.2",
- "mistralai/mistral-7b-v0.1",
- "mistralai/mixtral-8x7b-instruct-v0.1",
- // -------------------------------------
- // video model
- // -------------------------------------
- // "minimax/video-01", // TODO: implement the adaptor
+var RatioMap = map[string]ratio.Ratio{
+ "black-forest-labs/flux-1.1-pro": {Input: 0.04 * ratio.USD, Output: 0},
+ "black-forest-labs/flux-1.1-pro-ultra": {Input: 0.06 * ratio.USD, Output: 0},
+ "black-forest-labs/flux-canny-dev": {Input: 0.025 * ratio.USD, Output: 0},
+ "black-forest-labs/flux-canny-pro": {Input: 0.05 * ratio.USD, Output: 0},
+ "black-forest-labs/flux-depth-dev": {Input: 0.025 * ratio.USD, Output: 0},
+ "black-forest-labs/flux-depth-pro": {Input: 0.05 * ratio.USD, Output: 0},
+ "black-forest-labs/flux-dev": {Input: 0.025 * ratio.USD, Output: 0},
+ "black-forest-labs/flux-dev-lora": {Input: 0.032 * ratio.USD, Output: 0},
+ "black-forest-labs/flux-fill-dev": {Input: 0.04 * ratio.USD, Output: 0},
+ "black-forest-labs/flux-fill-pro": {Input: 0.05 * ratio.USD, Output: 0},
+ "black-forest-labs/flux-pro": {Input: 0.055 * ratio.USD, Output: 0},
+ "black-forest-labs/flux-redux-dev": {Input: 0.025 * ratio.USD, Output: 0},
+ "black-forest-labs/flux-redux-schnell": {Input: 0.003 * ratio.USD, Output: 0},
+ "black-forest-labs/flux-schnell": {Input: 0.003 * ratio.USD, Output: 0},
+ "black-forest-labs/flux-schnell-lora": {Input: 0.02 * ratio.USD, Output: 0},
+ "ideogram-ai/ideogram-v2": {Input: 0.08 * ratio.USD, Output: 0},
+ "ideogram-ai/ideogram-v2-turbo": {Input: 0.05 * ratio.USD, Output: 0},
+ "recraft-ai/recraft-v3": {Input: 0.04 * ratio.USD, Output: 0},
+ "recraft-ai/recraft-v3-svg": {Input: 0.08 * ratio.USD, Output: 0},
+ "stability-ai/stable-diffusion-3": {Input: 0.035 * ratio.USD, Output: 0},
+ "stability-ai/stable-diffusion-3.5-large": {Input: 0.065 * ratio.USD, Output: 0},
+ "stability-ai/stable-diffusion-3.5-large-turbo": {Input: 0.04 * ratio.USD, Output: 0},
+ "stability-ai/stable-diffusion-3.5-medium": {Input: 0.035 * ratio.USD, Output: 0},
+ // replicate chat models
+ "ibm-granite/granite-20b-code-instruct-8k": {Input: 0.100 * ratio.MILLI_USD, Output: 0.500 * ratio.MILLI_USD},
+ "ibm-granite/granite-3.0-2b-instruct": {Input: 0.030 * ratio.MILLI_USD, Output: 0.250 * ratio.MILLI_USD},
+ "ibm-granite/granite-3.0-8b-instruct": {Input: 0.050 * ratio.MILLI_USD, Output: 0.250 * ratio.MILLI_USD},
+ "ibm-granite/granite-8b-code-instruct-128k": {Input: 0.050 * ratio.MILLI_USD, Output: 0.250 * ratio.MILLI_USD},
+ "meta/llama-2-13b": {Input: 0.100 * ratio.MILLI_USD, Output: 0.500 * ratio.MILLI_USD},
+ "meta/llama-2-13b-chat": {Input: 0.100 * ratio.MILLI_USD, Output: 0.500 * ratio.MILLI_USD},
+ "meta/llama-2-70b": {Input: 0.650 * ratio.MILLI_USD, Output: 2.750 * ratio.MILLI_USD},
+ "meta/llama-2-70b-chat": {Input: 0.650 * ratio.MILLI_USD, Output: 2.750 * ratio.MILLI_USD},
+ "meta/llama-2-7b": {Input: 0.050 * ratio.MILLI_USD, Output: 0.250 * ratio.MILLI_USD},
+ "meta/llama-2-7b-chat": {Input: 0.050 * ratio.MILLI_USD, Output: 0.250 * ratio.MILLI_USD},
+ "meta/meta-llama-3.1-405b-instruct": {Input: 9.500 * ratio.MILLI_USD, Output: 9.500 * ratio.MILLI_USD},
+ "meta/meta-llama-3-70b": {Input: 0.650 * ratio.MILLI_USD, Output: 2.750 * ratio.MILLI_USD},
+ "meta/meta-llama-3-70b-instruct": {Input: 0.650 * ratio.MILLI_USD, Output: 2.750 * ratio.MILLI_USD},
+ "meta/meta-llama-3-8b": {Input: 0.050 * ratio.MILLI_USD, Output: 0.250 * ratio.MILLI_USD},
+ "meta/meta-llama-3-8b-instruct": {Input: 0.050 * ratio.MILLI_USD, Output: 0.250 * ratio.MILLI_USD},
+ "mistralai/mistral-7b-instruct-v0.2": {Input: 0.050 * ratio.MILLI_USD, Output: 0.250 * ratio.MILLI_USD},
+ "mistralai/mistral-7b-v0.1": {Input: 0.050 * ratio.MILLI_USD, Output: 0.250 * ratio.MILLI_USD},
+ "mistralai/mixtral-8x7b-instruct-v0.1": {Input: 0.300 * ratio.MILLI_USD, Output: 1.000 * ratio.MILLI_USD},
}
diff --git a/relay/adaptor/siliconflow/constants.go b/relay/adaptor/siliconflow/constants.go
index 0bf54761..59fb4e85 100644
--- a/relay/adaptor/siliconflow/constants.go
+++ b/relay/adaptor/siliconflow/constants.go
@@ -1,36 +1,40 @@
package siliconflow
-// https://docs.siliconflow.cn/docs/getting-started
+import "github.com/songquanpeng/one-api/relay/billing/ratio"
-var ModelList = []string{
- "deepseek-ai/deepseek-llm-67b-chat",
- "Qwen/Qwen1.5-14B-Chat",
- "Qwen/Qwen1.5-7B-Chat",
- "Qwen/Qwen1.5-110B-Chat",
- "Qwen/Qwen1.5-32B-Chat",
- "01-ai/Yi-1.5-6B-Chat",
- "01-ai/Yi-1.5-9B-Chat-16K",
- "01-ai/Yi-1.5-34B-Chat-16K",
- "THUDM/chatglm3-6b",
- "deepseek-ai/DeepSeek-V2-Chat",
- "THUDM/glm-4-9b-chat",
- "Qwen/Qwen2-72B-Instruct",
- "Qwen/Qwen2-7B-Instruct",
- "Qwen/Qwen2-57B-A14B-Instruct",
- "deepseek-ai/DeepSeek-Coder-V2-Instruct",
- "Qwen/Qwen2-1.5B-Instruct",
- "internlm/internlm2_5-7b-chat",
- "BAAI/bge-large-en-v1.5",
- "BAAI/bge-large-zh-v1.5",
- "Pro/Qwen/Qwen2-7B-Instruct",
- "Pro/Qwen/Qwen2-1.5B-Instruct",
- "Pro/Qwen/Qwen1.5-7B-Chat",
- "Pro/THUDM/glm-4-9b-chat",
- "Pro/THUDM/chatglm3-6b",
- "Pro/01-ai/Yi-1.5-9B-Chat-16K",
- "Pro/01-ai/Yi-1.5-6B-Chat",
- "Pro/google/gemma-2-9b-it",
- "Pro/internlm/internlm2_5-7b-chat",
- "Pro/meta-llama/Meta-Llama-3-8B-Instruct",
- "Pro/mistralai/Mistral-7B-Instruct-v0.2",
+// https://siliconflow.cn/zh-cn/models
+// https://siliconflow.cn/zh-cn/pricing
+var RatioMap = map[string]ratio.Ratio{
+ "Qwen/Qwen2.5-72B-Instruct": {Input: 41.3 * ratio.MILLI_RMB, Output: 41.3 * ratio.MILLI_RMB},
+ "Qwen/Qwen2.5-7B-Instruct": {Input: 3.5 * ratio.MILLI_RMB, Output: 3.5 * ratio.MILLI_RMB},
+ "deepseek-ai/deepseek-llm-67b-chat": {Input: 0, Output: 0},
+ "Qwen/Qwen1.5-14B-Chat": {Input: 0, Output: 0},
+ "Qwen/Qwen1.5-7B-Chat": {Input: 0, Output: 0},
+ "Qwen/Qwen1.5-110B-Chat": {Input: 0, Output: 0},
+ "Qwen/Qwen1.5-32B-Chat": {Input: 0, Output: 0},
+ "01-ai/Yi-1.5-6B-Chat": {Input: 0, Output: 0},
+ "01-ai/Yi-1.5-9B-Chat-16K": {Input: 0, Output: 0},
+ "01-ai/Yi-1.5-34B-Chat-16K": {Input: 0, Output: 0},
+ "THUDM/chatglm3-6b": {Input: 0, Output: 0},
+ "deepseek-ai/DeepSeek-V2-Chat": {Input: 0, Output: 0},
+ "THUDM/glm-4-9b-chat": {Input: 0, Output: 0},
+ "Qwen/Qwen2-72B-Instruct": {Input: 0, Output: 0},
+ "Qwen/Qwen2-7B-Instruct": {Input: 0, Output: 0},
+ "Qwen/Qwen2-57B-A14B-Instruct": {Input: 0, Output: 0},
+ "deepseek-ai/DeepSeek-Coder-V2-Instruct": {Input: 0, Output: 0},
+ "Qwen/Qwen2-1.5B-Instruct": {Input: 0, Output: 0},
+ "internlm/internlm2_5-7b-chat": {Input: 0, Output: 0},
+ "BAAI/bge-large-en-v1.5": {Input: 0, Output: 0},
+ "BAAI/bge-large-zh-v1.5": {Input: 0, Output: 0},
+ "Pro/Qwen/Qwen2-7B-Instruct": {Input: 0, Output: 0},
+ "Pro/Qwen/Qwen2-1.5B-Instruct": {Input: 0, Output: 0},
+ "Pro/Qwen/Qwen1.5-7B-Chat": {Input: 0, Output: 0},
+ "Pro/THUDM/glm-4-9b-chat": {Input: 0, Output: 0},
+ "Pro/THUDM/chatglm3-6b": {Input: 0, Output: 0},
+ "Pro/01-ai/Yi-1.5-9B-Chat-16K": {Input: 0, Output: 0},
+ "Pro/01-ai/Yi-1.5-6B-Chat": {Input: 0, Output: 0},
+ "Pro/google/gemma-2-9b-it": {Input: 0, Output: 0},
+ "Pro/internlm/internlm2_5-7b-chat": {Input: 0, Output: 0},
+ "Pro/meta-llama/Meta-Llama-3-8B-Instruct": {Input: 0, Output: 0},
+ "Pro/mistralai/Mistral-7B-Instruct-v0.2": {Input: 0, Output: 0},
}
diff --git a/relay/adaptor/stepfun/constants.go b/relay/adaptor/stepfun/constants.go
index 6a2346ca..b4b0aeda 100644
--- a/relay/adaptor/stepfun/constants.go
+++ b/relay/adaptor/stepfun/constants.go
@@ -1,13 +1,17 @@
package stepfun
-var ModelList = []string{
- "step-1-8k",
- "step-1-32k",
- "step-1-128k",
- "step-1-256k",
- "step-1-flash",
- "step-2-16k",
- "step-1v-8k",
- "step-1v-32k",
- "step-1x-medium",
+import "github.com/songquanpeng/one-api/relay/billing/ratio"
+
+// https://platform.stepfun.com/docs/pricing/details
+var RatioMap = map[string]ratio.Ratio{
+ "step-1-8k": {Input: 5 * ratio.MILLI_RMB, Output: 20 * ratio.MILLI_RMB},
+ "step-1-32k": {Input: 15 * ratio.MILLI_RMB, Output: 70 * ratio.MILLI_RMB},
+ "step-1-128k": {Input: 40 * ratio.MILLI_RMB, Output: 200 * ratio.MILLI_RMB},
+ "step-1-256k": {Input: 95 * ratio.MILLI_RMB, Output: 300 * ratio.MILLI_RMB},
+ "step-1-flash": {Input: 1 * ratio.MILLI_RMB, Output: 4 * ratio.MILLI_RMB},
+ "step-2-16k": {Input: 38 * ratio.MILLI_RMB, Output: 120 * ratio.MILLI_RMB},
+ "step-1v-8k": {Input: 5 * ratio.MILLI_RMB, Output: 20 * ratio.MILLI_RMB},
+ "step-1v-32k": {Input: 15 * ratio.MILLI_RMB, Output: 70 * ratio.MILLI_RMB},
+ "step-1.5v-mini": {Input: 8 * ratio.MILLI_RMB, Output: 35 * ratio.MILLI_RMB},
+ "step-1x-medium": {Input: 0.1 * ratio.RMB, Output: 0},
}
diff --git a/relay/adaptor/tencent/adaptor.go b/relay/adaptor/tencent/adaptor.go
index 0de92d4a..8c7eb507 100644
--- a/relay/adaptor/tencent/adaptor.go
+++ b/relay/adaptor/tencent/adaptor.go
@@ -2,16 +2,18 @@ package tencent
import (
"errors"
- "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/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/billing/ratio"
+ "github.com/songquanpeng/one-api/relay/meta"
+ "github.com/songquanpeng/one-api/relay/model"
)
// https://cloud.tencent.com/document/api/1729/101837
@@ -80,8 +82,12 @@ func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Met
return
}
+func (a *Adaptor) GetRatio(meta *meta.Meta) *ratio.Ratio {
+ return adaptor.GetRatioHelper(meta, RatioMap)
+}
+
func (a *Adaptor) GetModelList() []string {
- return ModelList
+ return adaptor.GetModelListHelper(RatioMap)
}
func (a *Adaptor) GetChannelName() string {
diff --git a/relay/adaptor/tencent/constants.go b/relay/adaptor/tencent/constants.go
index e8631e5f..ede8697d 100644
--- a/relay/adaptor/tencent/constants.go
+++ b/relay/adaptor/tencent/constants.go
@@ -1,9 +1,19 @@
package tencent
-var ModelList = []string{
- "hunyuan-lite",
- "hunyuan-standard",
- "hunyuan-standard-256K",
- "hunyuan-pro",
- "hunyuan-vision",
+import "github.com/songquanpeng/one-api/relay/billing/ratio"
+
+// https://cloud.tencent.com/document/product/1729/97731
+var RatioMap = map[string]ratio.Ratio{
+ "hunyuan-turbo": {Input: 0.015 * ratio.RMB, Output: 0.05 * ratio.RMB},
+ "hunyuan-large": {Input: 0.004 * ratio.RMB, Output: 0.012 * ratio.RMB},
+ "hunyuan-large-longcontext": {Input: 0.006 * ratio.RMB, Output: 0.018 * ratio.RMB},
+ "hunyuan-standard": {Input: 0.0008 * ratio.RMB, Output: 0.002 * ratio.RMB},
+ "hunyuan-standard-256K": {Input: 0.0005 * ratio.RMB, Output: 0.002 * ratio.RMB},
+ "hunyuan-translation-lite": {Input: 0.005 * ratio.RMB, Output: 0.015 * ratio.RMB},
+ "hunyuan-role": {Input: 0.004 * ratio.RMB, Output: 0.008 * ratio.RMB},
+ "hunyuan-functioncall": {Input: 0.004 * ratio.RMB, Output: 0.008 * ratio.RMB},
+ "hunyuan-code": {Input: 0.004 * ratio.RMB, Output: 0.008 * ratio.RMB},
+ "hunyuan-turbo-vision": {Input: 0.08 * ratio.RMB, Output: 0.08 * ratio.RMB},
+ "hunyuan-vision": {Input: 0.018 * ratio.RMB, Output: 0.018 * ratio.RMB},
+ "hunyuan-embedding": {Input: 0.0007 * ratio.RMB, Output: 0.0007 * ratio.RMB},
}
diff --git a/relay/adaptor/togetherai/constants.go b/relay/adaptor/togetherai/constants.go
index 0a79fbdc..5291122d 100644
--- a/relay/adaptor/togetherai/constants.go
+++ b/relay/adaptor/togetherai/constants.go
@@ -1,10 +1,12 @@
package togetherai
-// https://docs.together.ai/docs/inference-models
+import "github.com/songquanpeng/one-api/relay/billing/ratio"
-var ModelList = []string{
- "meta-llama/Llama-3-70b-chat-hf",
- "deepseek-ai/deepseek-coder-33b-instruct",
- "mistralai/Mixtral-8x22B-Instruct-v0.1",
- "Qwen/Qwen1.5-72B-Chat",
+// https://www.together.ai/pricing
+// https://docs.together.ai/docs/inference-models
+var RatioMap = map[string]ratio.Ratio{
+ "meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo": {Input: 0.88 * ratio.MILLI_USD, Output: 0.88 * ratio.MILLI_USD},
+ "deepseek-ai/deepseek-coder-33b-instruct": {Input: 1.25 * ratio.MILLI_USD, Output: 1.25 * ratio.MILLI_USD},
+ "mistralai/Mixtral-8x22B-Instruct-v0.1": {Input: 1.20 * ratio.MILLI_USD, Output: 1.20 * ratio.MILLI_USD},
+ "Qwen/Qwen2-72B-Instruct": {Input: 0.90 * ratio.MILLI_USD, Output: 0.90 * ratio.MILLI_USD},
}
diff --git a/relay/adaptor/vertexai/adaptor.go b/relay/adaptor/vertexai/adaptor.go
index 3fab4a45..b635a229 100644
--- a/relay/adaptor/vertexai/adaptor.go
+++ b/relay/adaptor/vertexai/adaptor.go
@@ -10,9 +10,9 @@ import (
"github.com/gin-gonic/gin"
"github.com/songquanpeng/one-api/relay/adaptor"
channelhelper "github.com/songquanpeng/one-api/relay/adaptor"
+ "github.com/songquanpeng/one-api/relay/billing/ratio"
"github.com/songquanpeng/one-api/relay/meta"
"github.com/songquanpeng/one-api/relay/model"
- relaymodel "github.com/songquanpeng/one-api/relay/model"
)
var _ adaptor.Adaptor = new(Adaptor)
@@ -40,19 +40,29 @@ func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.G
func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
adaptor := GetAdaptor(meta.ActualModelName)
if adaptor == nil {
- return nil, &relaymodel.ErrorWithStatusCode{
+ return nil, &model.ErrorWithStatusCode{
StatusCode: http.StatusInternalServerError,
- Error: relaymodel.Error{
- Message: "adaptor not found",
- },
+ Error: model.Error{Message: "adaptor not found"},
}
}
+
return adaptor.DoResponse(c, resp, meta)
}
-func (a *Adaptor) GetModelList() (models []string) {
- models = modelList
- return
+func (a *Adaptor) GetRatio(meta *meta.Meta) *ratio.Ratio {
+ adaptor := GetAdaptor(meta.ActualModelName)
+ if adaptor == nil {
+ return nil
+ }
+ return adaptor.GetRatio(meta)
+}
+
+func (a *Adaptor) GetModelList() []string {
+ var resp []string
+ for model := range modelMapping {
+ resp = append(resp, model)
+ }
+ return resp
}
func (a *Adaptor) GetChannelName() string {
diff --git a/relay/adaptor/vertexai/claude/adapter.go b/relay/adaptor/vertexai/claude/adapter.go
index cb911cfe..8f8288ff 100644
--- a/relay/adaptor/vertexai/claude/adapter.go
+++ b/relay/adaptor/vertexai/claude/adapter.go
@@ -6,19 +6,21 @@ import (
"github.com/gin-gonic/gin"
"github.com/pkg/errors"
"github.com/songquanpeng/one-api/common/ctxkey"
+ "github.com/songquanpeng/one-api/relay/adaptor"
"github.com/songquanpeng/one-api/relay/adaptor/anthropic"
-
+ "github.com/songquanpeng/one-api/relay/billing/ratio"
"github.com/songquanpeng/one-api/relay/meta"
"github.com/songquanpeng/one-api/relay/model"
)
-var ModelList = []string{
- "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",
+// https://cloud.google.com/vertex-ai/generative-ai/pricing?hl=zh-cn#claude-models
+var RatioMap = map[string]ratio.Ratio{
+ "claude-3-haiku@20240307": {Input: 0.25 * ratio.MILLI_USD, Output: 1.25 * ratio.MILLI_USD},
+ "claude-3-sonnet@20240229": {Input: 3 * ratio.MILLI_USD, Output: 15 * ratio.MILLI_USD},
+ "claude-3-opus@20240229": {Input: 15 * ratio.MILLI_USD, Output: 75 * ratio.MILLI_USD},
+ "claude-3-5-sonnet@20240620": {Input: 3 * ratio.MILLI_USD, Output: 15 * ratio.MILLI_USD},
+ "claude-3-5-sonnet-v2@20241022": {Input: 3 * ratio.MILLI_USD, Output: 15 * ratio.MILLI_USD},
+ "claude-3-5-haiku@20241022": {Input: 0.80 * ratio.MILLI_USD, Output: 4 * ratio.MILLI_USD},
}
const anthropicVersion = "vertex-2023-10-16"
@@ -58,3 +60,7 @@ func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Met
}
return
}
+
+func (a *Adaptor) GetRatio(meta *meta.Meta) *ratio.Ratio {
+ return adaptor.GetRatioHelper(meta, RatioMap)
+}
diff --git a/relay/adaptor/vertexai/gemini/adapter.go b/relay/adaptor/vertexai/gemini/adapter.go
index b5377875..27ac0d1b 100644
--- a/relay/adaptor/vertexai/gemini/adapter.go
+++ b/relay/adaptor/vertexai/gemini/adapter.go
@@ -6,20 +6,16 @@ import (
"github.com/gin-gonic/gin"
"github.com/pkg/errors"
"github.com/songquanpeng/one-api/common/ctxkey"
+ "github.com/songquanpeng/one-api/relay/adaptor"
"github.com/songquanpeng/one-api/relay/adaptor/gemini"
"github.com/songquanpeng/one-api/relay/adaptor/openai"
- "github.com/songquanpeng/one-api/relay/relaymode"
-
+ "github.com/songquanpeng/one-api/relay/billing/ratio"
"github.com/songquanpeng/one-api/relay/meta"
"github.com/songquanpeng/one-api/relay/model"
+ "github.com/songquanpeng/one-api/relay/relaymode"
)
-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",
-}
+var RatioMap = gemini.RatioMap
type Adaptor struct {
}
@@ -50,3 +46,7 @@ func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Met
}
return
}
+
+func (a *Adaptor) GetRatio(meta *meta.Meta) *ratio.Ratio {
+ return adaptor.GetRatioHelper(meta, RatioMap)
+}
diff --git a/relay/adaptor/vertexai/registry.go b/relay/adaptor/vertexai/registry.go
index 41099f02..966b91d5 100644
--- a/relay/adaptor/vertexai/registry.go
+++ b/relay/adaptor/vertexai/registry.go
@@ -6,6 +6,7 @@ import (
"github.com/gin-gonic/gin"
claude "github.com/songquanpeng/one-api/relay/adaptor/vertexai/claude"
gemini "github.com/songquanpeng/one-api/relay/adaptor/vertexai/gemini"
+ "github.com/songquanpeng/one-api/relay/billing/ratio"
"github.com/songquanpeng/one-api/relay/meta"
"github.com/songquanpeng/one-api/relay/model"
)
@@ -18,16 +19,13 @@ const (
)
var modelMapping = map[string]VertexAIModelType{}
-var modelList = []string{}
func init() {
- modelList = append(modelList, claude.ModelList...)
- for _, model := range claude.ModelList {
+ for model := range claude.RatioMap {
modelMapping[model] = VerterAIClaude
}
- modelList = append(modelList, gemini.ModelList...)
- for _, model := range gemini.ModelList {
+ for model := range gemini.RatioMap {
modelMapping[model] = VerterAIGemini
}
}
@@ -35,6 +33,7 @@ func init() {
type innerAIAdapter interface {
ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error)
DoResponse(c *gin.Context, resp *http.Response, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode)
+ GetRatio(meta *meta.Meta) *ratio.Ratio
}
func GetAdaptor(model string) innerAIAdapter {
diff --git a/relay/adaptor/xai/constants.go b/relay/adaptor/xai/constants.go
index 9082b999..ff720e44 100644
--- a/relay/adaptor/xai/constants.go
+++ b/relay/adaptor/xai/constants.go
@@ -1,5 +1,7 @@
package xai
-var ModelList = []string{
- "grok-beta",
+import "github.com/songquanpeng/one-api/relay/billing/ratio"
+
+var RatioMap = map[string]ratio.Ratio{
+ "grok-beta": {Input: 5.0 * ratio.MILLI_USD, Output: 15.0 * ratio.MILLI_USD},
}
diff --git a/relay/adaptor/xunfei/adaptor.go b/relay/adaptor/xunfei/adaptor.go
index b5967f26..ab45ecf2 100644
--- a/relay/adaptor/xunfei/adaptor.go
+++ b/relay/adaptor/xunfei/adaptor.go
@@ -2,14 +2,16 @@ package xunfei
import (
"errors"
- "github.com/gin-gonic/gin"
- "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"
"strings"
+
+ "github.com/gin-gonic/gin"
+ "github.com/songquanpeng/one-api/relay/adaptor"
+ "github.com/songquanpeng/one-api/relay/adaptor/openai"
+ "github.com/songquanpeng/one-api/relay/billing/ratio"
+ "github.com/songquanpeng/one-api/relay/meta"
+ "github.com/songquanpeng/one-api/relay/model"
)
type Adaptor struct {
@@ -77,8 +79,12 @@ func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Met
return
}
+func (a *Adaptor) GetRatio(meta *meta.Meta) *ratio.Ratio {
+ return adaptor.GetRatioHelper(meta, RatioMap)
+}
+
func (a *Adaptor) GetModelList() []string {
- return ModelList
+ return adaptor.GetModelListHelper(RatioMap)
}
func (a *Adaptor) GetChannelName() string {
diff --git a/relay/adaptor/xunfei/constants.go b/relay/adaptor/xunfei/constants.go
index 5b82ac29..2ac2253e 100644
--- a/relay/adaptor/xunfei/constants.go
+++ b/relay/adaptor/xunfei/constants.go
@@ -1,12 +1,14 @@
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",
+import "github.com/songquanpeng/one-api/relay/billing/ratio"
+
+var RatioMap = map[string]ratio.Ratio{
+ "SparkDesk": {Input: 1.2858, Output: 1.2858}, // ￥0.018 / 1k tokens
+ "SparkDesk-v1.1": {Input: 1.2858, Output: 1.2858}, // ￥0.018 / 1k tokens
+ "SparkDesk-v2.1": {Input: 1.2858, Output: 1.2858}, // ￥0.018 / 1k tokens
+ "SparkDesk-v3.1": {Input: 1.2858, Output: 1.2858}, // ￥0.018 / 1k tokens
+ "SparkDesk-v3.1-128K": {Input: 1.2858, Output: 1.2858}, // ￥0.018 / 1k tokens
+ "SparkDesk-v3.5": {Input: 1.2858, Output: 1.2858}, // ￥0.018 / 1k tokens
+ "SparkDesk-v3.5-32K": {Input: 1.2858, Output: 1.2858}, // ￥0.018 / 1k tokens
+ "SparkDesk-v4.0": {Input: 1.2858, Output: 1.2858}, // ￥0.018 / 1k tokens
}
diff --git a/relay/adaptor/zhipu/adaptor.go b/relay/adaptor/zhipu/adaptor.go
index 660bd379..863a491f 100644
--- a/relay/adaptor/zhipu/adaptor.go
+++ b/relay/adaptor/zhipu/adaptor.go
@@ -3,16 +3,18 @@ package zhipu
import (
"errors"
"fmt"
+ "io"
+ "net/http"
+ "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/billing/ratio"
"github.com/songquanpeng/one-api/relay/meta"
"github.com/songquanpeng/one-api/relay/model"
"github.com/songquanpeng/one-api/relay/relaymode"
- "io"
- "net/http"
- "strings"
)
type Adaptor struct {
@@ -140,8 +142,12 @@ func ConvertEmbeddingRequest(request model.GeneralOpenAIRequest) (*EmbeddingRequ
}, nil
}
+func (a *Adaptor) GetRatio(meta *meta.Meta) *ratio.Ratio {
+ return adaptor.GetRatioHelper(meta, RatioMap)
+}
+
func (a *Adaptor) GetModelList() []string {
- return ModelList
+ return adaptor.GetModelListHelper(RatioMap)
}
func (a *Adaptor) GetChannelName() string {
diff --git a/relay/adaptor/zhipu/constants.go b/relay/adaptor/zhipu/constants.go
index e1192123..b8fa39f9 100644
--- a/relay/adaptor/zhipu/constants.go
+++ b/relay/adaptor/zhipu/constants.go
@@ -1,7 +1,24 @@
package zhipu
-var ModelList = []string{
- "chatglm_turbo", "chatglm_pro", "chatglm_std", "chatglm_lite",
- "glm-4", "glm-4v", "glm-3-turbo", "embedding-2",
- "cogview-3",
+import "github.com/songquanpeng/one-api/relay/billing/ratio"
+
+var RatioMap = map[string]ratio.Ratio{
+ "glm-zero-preview": {Input: 0.01 * ratio.RMB, Output: 0.01 * ratio.RMB},
+ "glm-4-plus": {Input: 0.05 * ratio.RMB, Output: 0.05 * ratio.RMB},
+ "glm-4-0520": {Input: 0.1 * ratio.RMB, Output: 0.1 * ratio.RMB},
+ "glm-4-airx": {Input: 0.01 * ratio.RMB, Output: 0.01 * ratio.RMB},
+ "glm-4-air": {Input: 0.0005 * ratio.RMB, Output: 0.0005 * ratio.RMB},
+ "glm-4-long": {Input: 0.001 * ratio.RMB, Output: 0.001 * ratio.RMB},
+ "glm-4-flashx": {Input: 0.0001 * ratio.RMB, Output: 0.0001 * ratio.RMB},
+ "glm-4v-plus": {Input: 0.004 * ratio.RMB, Output: 0.004 * ratio.RMB},
+ "glm-4v": {Input: 0.05 * ratio.RMB, Output: 0},
+ "cogview-3-plus": {Input: 0.06 * ratio.RMB, Output: 0},
+ "cogview-3": {Input: 0.1 * ratio.RMB, Output: 0},
+ "cogvideox": {Input: 0.5 * ratio.RMB, Output: 0},
+ "embedding-3": {Input: 0.0005 * ratio.RMB, Output: 0},
+ "embedding-2": {Input: 0.0005 * ratio.RMB, Output: 0},
+ "glm-4-flash": {Input: 0, Output: 0}, // 免费
+ "glm-4v-flash": {Input: 0, Output: 0}, // 免费
+ "cogview-3-flash": {Input: 0, Output: 0}, // 免费
+ "cogvideox-flash": {Input: 0, Output: 0}, // 免费
}
diff --git a/relay/adaptor/zhipu/main.go b/relay/adaptor/zhipu/main.go
index ab3a5678..488a0df1 100644
--- a/relay/adaptor/zhipu/main.go
+++ b/relay/adaptor/zhipu/main.go
@@ -3,13 +3,14 @@ package zhipu
import (
"bufio"
"encoding/json"
- "github.com/songquanpeng/one-api/common/render"
"io"
"net/http"
"strings"
"sync"
"time"
+ "github.com/songquanpeng/one-api/common/render"
+
"github.com/gin-gonic/gin"
"github.com/golang-jwt/jwt"
"github.com/songquanpeng/one-api/common"
@@ -268,6 +269,9 @@ func EmbeddingsHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithSta
c.Writer.Header().Set("Content-Type", "application/json")
c.Writer.WriteHeader(resp.StatusCode)
_, err = c.Writer.Write(jsonResponse)
+ if err != nil {
+ return openai.ErrorWrapper(err, "write_response_body_failed", http.StatusInternalServerError), nil
+ }
return nil, &fullTextResponse.Usage
}
@@ -276,11 +280,7 @@ func embeddingResponseZhipu2OpenAI(response *EmbeddingResponse) *openai.Embeddin
Object: "list",
Data: make([]openai.EmbeddingResponseItem, 0, len(response.Embeddings)),
Model: response.Model,
- Usage: model.Usage{
- PromptTokens: response.PromptTokens,
- CompletionTokens: response.CompletionTokens,
- TotalTokens: response.Usage.TotalTokens,
- },
+ Usage: response.Usage,
}
for _, item := range response.Embeddings {
diff --git a/relay/billing/billing.go b/relay/billing/billing.go
index a99d37ee..0a66bacd 100644
--- a/relay/billing/billing.go
+++ b/relay/billing/billing.go
@@ -3,6 +3,7 @@ package billing
import (
"context"
"fmt"
+
"github.com/songquanpeng/one-api/common/logger"
"github.com/songquanpeng/one-api/model"
)
@@ -19,20 +20,22 @@ func ReturnPreConsumedQuota(ctx context.Context, preConsumedQuota int64, tokenId
}
}
-func PostConsumeQuota(ctx context.Context, tokenId int, quotaDelta int64, totalQuota int64, userId int, channelId int, modelRatio float64, groupRatio float64, modelName string, tokenName string) {
+func PostConsumeQuota(ctx context.Context, tokenId int, quotaDelta int64, totalQuota int64, userId int, channelId int, modelRatio float64, groupRatio float64, modelName string, tokenName string, promptTokens int, completionTokens int) {
// quotaDelta is remaining quota to be consumed
- err := model.PostConsumeTokenQuota(tokenId, quotaDelta)
- if err != nil {
- logger.SysError("error consuming token remain quota: " + err.Error())
+ if quotaDelta != 0 {
+ err := model.PostConsumeTokenQuota(tokenId, quotaDelta)
+ if err != nil {
+ logger.SysError("error consuming token remain quota: " + err.Error())
+ }
}
- err = model.CacheUpdateUserQuota(ctx, userId)
+ err := model.CacheUpdateUserQuota(ctx, userId)
if err != nil {
logger.SysError("error update user quota cache: " + err.Error())
}
// 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)
+ model.RecordConsumeLog(ctx, userId, channelId, promptTokens, completionTokens, modelName, tokenName, totalQuota, logContent)
model.UpdateUserUsedQuotaAndRequestCount(userId, totalQuota)
model.UpdateChannelUsedQuota(channelId, totalQuota)
}
diff --git a/relay/billing/ratio/model.go b/relay/billing/ratio/model.go
index f83aa70c..80dad948 100644
--- a/relay/billing/ratio/model.go
+++ b/relay/billing/ratio/model.go
@@ -9,17 +9,35 @@ import (
)
const (
- USD2RMB = 7
- USD = 500 // $0.002 = 1 -> $1 = 500
- RMB = USD / USD2RMB
+ USD2RMB = 7
+ USD = 500 // $0.002 = 1 -> $1 = 500
+ RMB = USD / USD2RMB // 1RMB = 1/7USD
+ MILLI_USD = 1.0 / 1000 * USD
+ MILLI_RMB = 1.0 / 1000 * RMB
+ TokensPerSecond = 1000 / 20 // $0.006 / minute -> $0.002 / 20 seconds -> $0.002 / 1K tokens
)
+type Ratio struct {
+ Input float64 `json:"input,omitempty"` // input ratio
+ Output float64 `json:"output,omitempty"` // output ratio
+ LongThreshold int `json:"long_threshold,omitempty"` // for gemini like models, prompt longer than threshold will be charged as long input
+ LongInput float64 `json:"long_input,omitempty"` // long input ratio
+ LongOutput float64 `json:"long_output,omitempty"` // long output ratio
+}
+
+var (
+ FallbackRatio = Ratio{Input: 30, Output: 30}
+)
+
+// Deprecated
+// TODO: remove this
// ModelRatio
// https://platform.openai.com/docs/models/model-endpoint-compatibility
// https://cloud.baidu.com/doc/WENXINWORKSHOP/s/Blfmc9dlf
// https://openai.com/pricing
// 1 === $0.002 / 1K tokens
// 1 === ￥0.014 / 1k tokens
+// 1 === $0.002 / 20 seconds (50 tokens per second)
var ModelRatio = map[string]float64{
// https://openai.com/pricing
"gpt-4": 15,
@@ -342,6 +360,7 @@ var CompletionRatio = map[string]float64{
var (
DefaultModelRatio map[string]float64
DefaultCompletionRatio map[string]float64
+ DefaultRatio = make(map[string]Ratio)
)
func init() {
@@ -536,3 +555,30 @@ func GetCompletionRatio(name string, channelType int) float64 {
return 1
}
+
+func Ratio2JSONString() string {
+ jsonBytes, err := json.Marshal(DefaultRatio)
+ if err != nil {
+ logger.SysError("error marshalling ratio: " + err.Error())
+ }
+ return string(jsonBytes)
+}
+
+func UpdateRatioByJSONString(jsonStr string) error {
+ DefaultRatio = make(map[string]Ratio)
+ return json.Unmarshal([]byte(jsonStr), &DefaultRatio)
+}
+
+func GetRatio(name string, channelType int) *Ratio {
+ var result Ratio
+ model := fmt.Sprintf("%s(%d)", name, channelType)
+ if ratio, ok := DefaultRatio[model]; ok {
+ result = ratio
+ return &result
+ }
+ if ratio, ok := DefaultRatio[name]; ok {
+ result = ratio
+ return &result
+ }
+ return nil
+}
diff --git a/relay/controller/audio.go b/relay/controller/audio.go
index e3d57b1e..7d076279 100644
--- a/relay/controller/audio.go
+++ b/relay/controller/audio.go
@@ -18,6 +18,7 @@ import (
"github.com/songquanpeng/one-api/common/ctxkey"
"github.com/songquanpeng/one-api/common/logger"
"github.com/songquanpeng/one-api/model"
+ "github.com/songquanpeng/one-api/relay"
"github.com/songquanpeng/one-api/relay/adaptor/openai"
"github.com/songquanpeng/one-api/relay/billing"
billingratio "github.com/songquanpeng/one-api/relay/billing/ratio"
@@ -54,9 +55,16 @@ func RelayAudioHelper(c *gin.Context, relayMode int) *relaymodel.ErrorWithStatus
}
}
- modelRatio := billingratio.GetModelRatio(audioModel, channelType)
+ adaptor := relay.GetAdaptor(meta.APIType)
+ if adaptor == nil {
+ return openai.ErrorWrapper(fmt.Errorf("invalid api type: %d", meta.APIType), "invalid_api_type", http.StatusBadRequest)
+ }
+ adaptor.Init(meta)
+
groupRatio := billingratio.GetGroupRatio(group)
- ratio := modelRatio * groupRatio
+ adaptorRatio := GetRatio(meta, adaptor)
+ ratio := adaptorRatio.Input * groupRatio
+
var quota int64
var preConsumedQuota int64
switch relayMode {
@@ -216,7 +224,7 @@ func RelayAudioHelper(c *gin.Context, relayMode int) *relaymodel.ErrorWithStatus
succeed = true
quotaDelta := quota - preConsumedQuota
defer func(ctx context.Context) {
- go billing.PostConsumeQuota(ctx, tokenId, quotaDelta, quota, userId, channelId, modelRatio, groupRatio, audioModel, tokenName)
+ go billing.PostConsumeQuota(ctx, tokenId, quotaDelta, quota, userId, channelId, adaptorRatio.Input, groupRatio, audioModel, tokenName, 0, 0)
}(c.Request.Context())
for k, v := range resp.Header {
diff --git a/relay/controller/helper.go b/relay/controller/helper.go
index 5f5fc90c..1a76baa7 100644
--- a/relay/controller/helper.go
+++ b/relay/controller/helper.go
@@ -4,7 +4,6 @@ import (
"context"
"errors"
"fmt"
- "github.com/songquanpeng/one-api/relay/constant/role"
"math"
"net/http"
"strings"
@@ -14,9 +13,12 @@ import (
"github.com/songquanpeng/one-api/common/config"
"github.com/songquanpeng/one-api/common/logger"
"github.com/songquanpeng/one-api/model"
+ "github.com/songquanpeng/one-api/relay/adaptor"
"github.com/songquanpeng/one-api/relay/adaptor/openai"
+ "github.com/songquanpeng/one-api/relay/billing/ratio"
billingratio "github.com/songquanpeng/one-api/relay/billing/ratio"
"github.com/songquanpeng/one-api/relay/channeltype"
+ "github.com/songquanpeng/one-api/relay/constant/role"
"github.com/songquanpeng/one-api/relay/controller/validator"
"github.com/songquanpeng/one-api/relay/meta"
relaymodel "github.com/songquanpeng/one-api/relay/model"
@@ -91,17 +93,26 @@ func preConsumeQuota(ctx context.Context, textRequest *relaymodel.GeneralOpenAIR
return preConsumedQuota, nil
}
-func postConsumeQuota(ctx context.Context, usage *relaymodel.Usage, meta *meta.Meta, textRequest *relaymodel.GeneralOpenAIRequest, ratio float64, preConsumedQuota int64, modelRatio float64, groupRatio float64, systemPromptReset bool) {
+func postConsumeQuota(ctx context.Context, usage *relaymodel.Usage, meta *meta.Meta, textRequest *relaymodel.GeneralOpenAIRequest, ratio billingratio.Ratio, preConsumedQuota int64, groupRatio float64, systemPromptReset bool) {
if usage == nil {
logger.Error(ctx, "usage is nil, which is unexpected")
return
}
var quota int64
- completionRatio := billingratio.GetCompletionRatio(textRequest.Model, meta.ChannelType)
+ // use meta.OriginalModelName instead of mapped model name, which may named randomly in azure
promptTokens := usage.PromptTokens
completionTokens := usage.CompletionTokens
- quota = int64(math.Ceil((float64(promptTokens) + float64(completionTokens)*completionRatio) * ratio))
- if ratio != 0 && quota <= 0 {
+ promptRatio := ratio.Input
+ completionRatio := ratio.Output
+
+ // for gemini, prompt longer than 128k will be charged as long input
+ if ratio.LongInput > 0 && promptTokens > ratio.LongThreshold {
+ promptRatio = ratio.LongInput
+ completionRatio = ratio.LongOutput
+ }
+ quota = int64(math.Ceil(groupRatio * (float64(promptTokens)*promptRatio + float64(completionTokens)*completionRatio)))
+
+ if quota <= 0 && (ratio.Input > 0 || ratio.Output > 0) {
quota = 1
}
totalTokens := promptTokens + completionTokens
@@ -123,8 +134,8 @@ func postConsumeQuota(ctx context.Context, usage *relaymodel.Usage, meta *meta.M
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("模型倍率 %.2f，分组倍率 %.2f，补全倍率 %.2f%s", promptRatio, groupRatio, completionRatio/promptRatio, extraLog)
+ model.RecordConsumeLog(ctx, meta.UserId, meta.ChannelId, promptTokens, completionTokens, meta.OriginModelName, meta.TokenName, quota, logContent)
model.UpdateUserUsedQuotaAndRequestCount(meta.UserId, quota)
model.UpdateChannelUsedQuota(meta.ChannelId, quota)
}
@@ -185,3 +196,16 @@ func setSystemPrompt(ctx context.Context, request *relaymodel.GeneralOpenAIReque
logger.Infof(ctx, "add system prompt")
return true
}
+
+func GetRatio(meta *meta.Meta, adaptor adaptor.Adaptor) ratio.Ratio {
+ result := billingratio.GetRatio(meta.OriginModelName, meta.ChannelType)
+ if result != nil {
+ return *result
+ }
+ ratio := adaptor.GetRatio(meta)
+ if ratio != nil {
+ return *ratio
+ }
+ logger.SysError("model ratio not found: " + meta.OriginModelName)
+ return billingratio.FallbackRatio
+}
diff --git a/relay/controller/image.go b/relay/controller/image.go
index 1b69d97d..94d182bb 100644
--- a/relay/controller/image.go
+++ b/relay/controller/image.go
@@ -128,7 +128,6 @@ func RelayImageHelper(c *gin.Context, relayMode int) *relaymodel.ErrorWithStatus
return openai.ErrorWrapper(err, "get_image_cost_ratio_failed", http.StatusInternalServerError)
}
- imageModel := imageRequest.Model
// Convert the original image model
imageRequest.Model, _ = getMappedModelName(imageRequest.Model, billingratio.ImageOriginModelName)
c.Set("response_format", imageRequest.ResponseFormat)
@@ -167,9 +166,9 @@ func RelayImageHelper(c *gin.Context, relayMode int) *relaymodel.ErrorWithStatus
requestBody = bytes.NewBuffer(jsonStr)
}
- modelRatio := billingratio.GetModelRatio(imageModel, meta.ChannelType)
groupRatio := billingratio.GetGroupRatio(meta.Group)
- ratio := modelRatio * groupRatio
+ adaptorRatio := GetRatio(meta, adaptor)
+ ratio := adaptorRatio.Input * groupRatio
userQuota, err := model.CacheGetUserQuota(ctx, meta.UserId)
var quota int64
@@ -209,7 +208,7 @@ 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)
+ logContent := fmt.Sprintf("模型倍率 %.2f，分组倍率 %.2f", adaptorRatio.Input, groupRatio)
model.RecordConsumeLog(ctx, meta.UserId, meta.ChannelId, 0, 0, imageRequest.Model, tokenName, quota, logContent)
model.UpdateUserUsedQuotaAndRequestCount(meta.UserId, quota)
channelId := c.GetInt(ctxkey.ChannelId)
diff --git a/relay/controller/text.go b/relay/controller/text.go
index 9a47c58b..9e8a6f9d 100644
--- a/relay/controller/text.go
+++ b/relay/controller/text.go
@@ -4,10 +4,11 @@ import (
"bytes"
"encoding/json"
"fmt"
- "github.com/songquanpeng/one-api/common/config"
"io"
"net/http"
+ "github.com/songquanpeng/one-api/common/config"
+
"github.com/gin-gonic/gin"
"github.com/songquanpeng/one-api/common/logger"
"github.com/songquanpeng/one-api/relay"
@@ -32,6 +33,12 @@ func RelayTextHelper(c *gin.Context) *model.ErrorWithStatusCode {
}
meta.IsStream = textRequest.Stream
+ adaptor := relay.GetAdaptor(meta.APIType)
+ if adaptor == nil {
+ return openai.ErrorWrapper(fmt.Errorf("invalid api type: %d", meta.APIType), "invalid_api_type", http.StatusBadRequest)
+ }
+ adaptor.Init(meta)
+
// map model name
meta.OriginModelName = textRequest.Model
textRequest.Model, _ = getMappedModelName(textRequest.Model, meta.ModelMapping)
@@ -39,9 +46,10 @@ func RelayTextHelper(c *gin.Context) *model.ErrorWithStatusCode {
// set system prompt if not empty
systemPromptReset := setSystemPrompt(ctx, textRequest, meta.SystemPrompt)
// get model ratio & group ratio
- modelRatio := billingratio.GetModelRatio(textRequest.Model, meta.ChannelType)
groupRatio := billingratio.GetGroupRatio(meta.Group)
- ratio := modelRatio * groupRatio
+ adaptorRatio := GetRatio(meta, adaptor)
+ ratio := adaptorRatio.Input * groupRatio
+
// pre-consume quota
promptTokens := getPromptTokens(textRequest, meta.Mode)
meta.PromptTokens = promptTokens
@@ -51,12 +59,6 @@ func RelayTextHelper(c *gin.Context) *model.ErrorWithStatusCode {
return bizErr
}
- adaptor := relay.GetAdaptor(meta.APIType)
- if adaptor == nil {
- return openai.ErrorWrapper(fmt.Errorf("invalid api type: %d", meta.APIType), "invalid_api_type", http.StatusBadRequest)
- }
- adaptor.Init(meta)
-
// get request body
requestBody, err := getRequestBody(c, meta, textRequest, adaptor)
if err != nil {
@@ -82,7 +84,7 @@ func RelayTextHelper(c *gin.Context) *model.ErrorWithStatusCode {
return respErr
}
// post-consume quota
- go postConsumeQuota(ctx, usage, meta, textRequest, ratio, preConsumedQuota, modelRatio, groupRatio, systemPromptReset)
+ go postConsumeQuota(ctx, usage, meta, textRequest, adaptorRatio, preConsumedQuota, groupRatio, systemPromptReset)
return nil
}
diff --git a/web/default/src/components/OperationSetting.js b/web/default/src/components/OperationSetting.js
index 6356ac66..88cbf2d0 100644
--- a/web/default/src/components/OperationSetting.js
+++ b/web/default/src/components/OperationSetting.js
@@ -2,6 +2,12 @@ import React, { useEffect, useState } from 'react';
import { Divider, Form, Grid, Header } from 'semantic-ui-react';
import { API, showError, showSuccess, timestamp2string, verifyJSON } from '../helpers';
+const RATIO_MAPPING_EXAMPLE = {
+ 'gpt-4o-mini': {'input': 0.075, 'output': 0.3},
+ 'llama3-8b-8192(33)': {'input': 0.15, 'output': 0.3},
+ 'llama3-70b-8192(33)': {'input': 1.325, 'output': 1.749},
+};
+
const OperationSetting = () => {
let now = new Date();
let [inputs, setInputs] = useState({
@@ -10,9 +16,10 @@ const OperationSetting = () => {
QuotaForInvitee: 0,
QuotaRemindThreshold: 0,
PreConsumedQuota: 0,
- ModelRatio: '',
- CompletionRatio: '',
+ ModelRatio: '', // Deprecated
+ CompletionRatio: '', // Deprecated
GroupRatio: '',
+ Ratio: '',
TopUpLink: '',
ChatLink: '',
QuotaPerUnit: 0,
@@ -35,7 +42,7 @@ const OperationSetting = () => {
if (success) {
let newInputs = {};
data.forEach((item) => {
- if (item.key === 'ModelRatio' || item.key === 'GroupRatio' || item.key === 'CompletionRatio') {
+ if (item.key === 'ModelRatio' || item.key === 'GroupRatio' || item.key === 'CompletionRatio' || item.key === 'Ratio') {
item.value = JSON.stringify(JSON.parse(item.value), null, 2);
}
if (item.value === '{}') {
@@ -112,6 +119,13 @@ const OperationSetting = () => {
}
await updateOption('CompletionRatio', inputs.CompletionRatio);
}
+ if (originInputs['Ratio'] !== inputs.Ratio) {
+ if (!verifyJSON(inputs.Ratio)) {
+ showError('倍率不是合法的 JSON 字符串');
+ return;
+ }
+ await updateOption('Ratio', inputs.Ratio);
+ }
break;
case 'quota':
if (originInputs['QuotaForNewUser'] !== inputs.QuotaForNewUser) {
@@ -346,7 +360,18 @@ const OperationSetting = () => {
+
+
+ {
Date: Fri, 17 Jan 2025 15:12:14 +0800
Subject: [PATCH 2/6] fix: whisper model
---
common/ctxkey/key.go | 1 -
common/gin.go | 10 +++++-----
2 files changed, 5 insertions(+), 6 deletions(-)
diff --git a/common/ctxkey/key.go b/common/ctxkey/key.go
index 115558a5..681e2344 100644
--- a/common/ctxkey/key.go
+++ b/common/ctxkey/key.go
@@ -19,6 +19,5 @@ const (
TokenName = "token_name"
BaseURL = "base_url"
AvailableModels = "available_models"
- KeyRequestBody = "key_request_body"
SystemPrompt = "system_prompt"
)
diff --git a/common/gin.go b/common/gin.go
index 815b4ee5..1f5849b2 100644
--- a/common/gin.go
+++ b/common/gin.go
@@ -3,14 +3,14 @@ package common
import (
"bytes"
"encoding/json"
- "github.com/gin-gonic/gin"
- "github.com/songquanpeng/one-api/common/ctxkey"
"io"
"strings"
+
+ "github.com/gin-gonic/gin"
)
func GetRequestBody(c *gin.Context) ([]byte, error) {
- requestBody, _ := c.Get(ctxkey.KeyRequestBody)
+ requestBody, _ := c.Get(gin.BodyBytesKey)
if requestBody != nil {
return requestBody.([]byte), nil
}
@@ -19,7 +19,7 @@ func GetRequestBody(c *gin.Context) ([]byte, error) {
return nil, err
}
_ = c.Request.Body.Close()
- c.Set(ctxkey.KeyRequestBody, requestBody)
+ c.Set(gin.BodyBytesKey, requestBody)
return requestBody.([]byte), nil
}
@@ -31,7 +31,6 @@ func UnmarshalBodyReusable(c *gin.Context, v any) error {
contentType := c.Request.Header.Get("Content-Type")
if strings.HasPrefix(contentType, "application/json") {
err = json.Unmarshal(requestBody, &v)
- c.Request.Body = io.NopCloser(bytes.NewBuffer(requestBody))
} else {
c.Request.Body = io.NopCloser(bytes.NewBuffer(requestBody))
err = c.ShouldBind(&v)
@@ -40,6 +39,7 @@ func UnmarshalBodyReusable(c *gin.Context, v any) error {
return err
}
// Reset request body
+ c.Request.Body = io.NopCloser(bytes.NewBuffer(requestBody))
return nil
}
From 95b4967f40b44d254e6e5a8ac1d4f9938f2c9f60 Mon Sep 17 00:00:00 2001
From: WqyJh <781345688@qq.com>
Date: Fri, 17 Jan 2025 15:30:34 +0800
Subject: [PATCH 3/6] fix: ratio description
---
relay/adaptor/openai/constants.go | 8 ++++----
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/relay/adaptor/openai/constants.go b/relay/adaptor/openai/constants.go
index f749e1e6..ff243c0b 100644
--- a/relay/adaptor/openai/constants.go
+++ b/relay/adaptor/openai/constants.go
@@ -52,10 +52,10 @@ var RatioMap = map[string]ratio.Ratio{
"text-embedding-3-small": {Input: 0.01, Output: 0}, // $0.0002 / 1K tokens
"text-embedding-3-large": {Input: 0.065, Output: 0}, // $0.0013 / 1K tokens
"text-search-ada-doc-001": {Input: 10, Output: 0}, // $0.02 / 1K tokens
- "text-moderation-stable": {Input: 0.1, Output: 0}, // $0.002 / 1K tokens
- "text-moderation-latest": {Input: 0.1, Output: 0}, // $0.002 / 1K tokens
- "omni-moderation-latest": {Input: 0.1, Output: 0}, // $0.002 / 1K tokens
- "omni-moderation-2024-09-26": {Input: 0.1, Output: 0}, // $0.002 / 1K tokens
+ "text-moderation-stable": {Input: 0.1, Output: 0}, // currently free to use
+ "text-moderation-latest": {Input: 0.1, Output: 0}, // currently free to use
+ "omni-moderation-latest": {Input: 0.1, Output: 0}, // currently free to use
+ "omni-moderation-2024-09-26": {Input: 0.1, Output: 0}, // currently free to use
"dall-e-2": {Input: 0.02 * ratio.USD, Output: 0}, // $0.016 - $0.020 / image
"dall-e-3": {Input: 0.04 * ratio.USD, Output: 0}, // $0.040 - $0.120 / image
}
From 215e59b76d90649d6b9693dea902507e1b67e45f Mon Sep 17 00:00:00 2001
From: WqyJh <781345688@qq.com>
Date: Fri, 17 Jan 2025 16:06:04 +0800
Subject: [PATCH 4/6] fix: use gin.BodyBytesKey as KeyRequestBody
---
common/ctxkey/key.go | 3 +++
common/gin.go | 5 +++--
2 files changed, 6 insertions(+), 2 deletions(-)
diff --git a/common/ctxkey/key.go b/common/ctxkey/key.go
index 681e2344..e78e9d44 100644
--- a/common/ctxkey/key.go
+++ b/common/ctxkey/key.go
@@ -1,5 +1,7 @@
package ctxkey
+import "github.com/gin-gonic/gin"
+
const (
Config = "config"
Id = "id"
@@ -19,5 +21,6 @@ const (
TokenName = "token_name"
BaseURL = "base_url"
AvailableModels = "available_models"
+ KeyRequestBody = gin.BodyBytesKey
SystemPrompt = "system_prompt"
)
diff --git a/common/gin.go b/common/gin.go
index 1f5849b2..e3281fee 100644
--- a/common/gin.go
+++ b/common/gin.go
@@ -7,10 +7,11 @@ import (
"strings"
"github.com/gin-gonic/gin"
+ "github.com/songquanpeng/one-api/common/ctxkey"
)
func GetRequestBody(c *gin.Context) ([]byte, error) {
- requestBody, _ := c.Get(gin.BodyBytesKey)
+ requestBody, _ := c.Get(ctxkey.KeyRequestBody)
if requestBody != nil {
return requestBody.([]byte), nil
}
@@ -19,7 +20,7 @@ func GetRequestBody(c *gin.Context) ([]byte, error) {
return nil, err
}
_ = c.Request.Body.Close()
- c.Set(gin.BodyBytesKey, requestBody)
+ c.Set(ctxkey.KeyRequestBody, requestBody)
return requestBody.([]byte), nil
}
From 56df7c097a2cc060cb2b43b3668eafed8295c4f8 Mon Sep 17 00:00:00 2001
From: WqyJh <781345688@qq.com>
Date: Mon, 20 Jan 2025 17:51:26 +0800
Subject: [PATCH 5/6] feat: atomatically migrate old ratio to new
---
model/option.go | 20 +++++++++++
relay/billing/ratio/model.go | 65 ++++++++++++++++++++++++++++++++++++
2 files changed, 85 insertions(+)
diff --git a/model/option.go b/model/option.go
index c749b231..030f0b54 100644
--- a/model/option.go
+++ b/model/option.go
@@ -83,15 +83,35 @@ func InitOptionMap() {
func loadOptionsFromDatabase() {
options, _ := AllOption()
+ var oldModelRatio string
+ var oldCompletionRatio string
for _, option := range options {
if option.Key == "ModelRatio" {
+ oldModelRatio = option.Value
option.Value = billingratio.AddNewMissingRatio(option.Value)
}
+ if option.Key == "CompletionRatio" {
+ oldCompletionRatio = option.Value
+ }
err := updateOptionMap(option.Key, option.Value)
if err != nil {
logger.SysError("failed to update option map: " + err.Error())
}
}
+ for _, option := range options {
+ if option.Key == "Ratio" {
+ option.Value = billingratio.AddOldRatio(oldModelRatio, oldCompletionRatio)
+ err := updateOptionMap(option.Key, option.Value)
+ if err != nil {
+ logger.SysError("failed to update option map: " + err.Error())
+ }
+ err = UpdateOption(option.Key, option.Value)
+ if err != nil {
+ logger.SysError("failed to update option map: " + err.Error())
+ }
+ logger.SysLog("ratio merged")
+ }
+ }
}
func SyncOptions(frequency int) {
diff --git a/relay/billing/ratio/model.go b/relay/billing/ratio/model.go
index 80dad948..da795490 100644
--- a/relay/billing/ratio/model.go
+++ b/relay/billing/ratio/model.go
@@ -3,6 +3,7 @@ package ratio
import (
"encoding/json"
"fmt"
+ "strconv"
"strings"
"github.com/songquanpeng/one-api/common/logger"
@@ -394,6 +395,58 @@ func AddNewMissingRatio(oldRatio string) string {
return string(jsonBytes)
}
+func AddOldRatio(oldRatio string, oldCompletionRatio string) string {
+ modelRatio := make(map[string]float64)
+ if oldRatio != "" {
+ err := json.Unmarshal([]byte(oldRatio), &modelRatio)
+ if err != nil {
+ logger.SysError("error unmarshalling old ratio: " + err.Error())
+ return oldRatio
+ }
+ }
+
+ completionRatio := make(map[string]float64)
+ if oldCompletionRatio != "" {
+ err := json.Unmarshal([]byte(oldCompletionRatio), &completionRatio)
+ if err != nil {
+ logger.SysError("error unmarshalling old completion ratio: " + err.Error())
+ return oldCompletionRatio
+ }
+ }
+
+ newRatio := make(map[string]Ratio)
+
+ for k, v := range DefaultRatio {
+ if _, ok := newRatio[k]; !ok {
+ newRatio[k] = v
+ }
+ }
+
+ for k, v := range modelRatio {
+ if _, ok := DefaultRatio[k]; ok {
+ continue
+ }
+ modelName, channelType := SplitModelName(k)
+ ratio := Ratio{}
+ ratio.Input = v
+
+ if val, ok := completionRatio[k]; ok {
+ ratio.Output = v * val
+ } else {
+ ratio.Output = v * GetCompletionRatio(modelName, channelType)
+ }
+
+ newRatio[k] = ratio
+ }
+
+ jsonBytes, err := json.Marshal(newRatio)
+ if err != nil {
+ logger.SysError("error marshalling new ratio: " + err.Error())
+ return oldRatio
+ }
+ return string(jsonBytes)
+}
+
func ModelRatio2JSONString() string {
jsonBytes, err := json.Marshal(ModelRatio)
if err != nil {
@@ -444,6 +497,18 @@ func UpdateCompletionRatioByJSONString(jsonStr string) error {
return json.Unmarshal([]byte(jsonStr), &CompletionRatio)
}
+func SplitModelName(name string) (string, int) {
+ model := strings.Split(name, "(")
+ modelName := model[0]
+ channelType := 0
+ if len(model) > 1 {
+ if v, err := strconv.Atoi(model[1]); err == nil {
+ channelType = v
+ }
+ }
+ return modelName, channelType
+}
+
func GetCompletionRatio(name string, channelType int) float64 {
if strings.HasPrefix(name, "qwen-") && strings.HasSuffix(name, "-internet") {
name = strings.TrimSuffix(name, "-internet")
From f2bff58498906aec3c7b8660d6cd8577d31f1884 Mon Sep 17 00:00:00 2001
From: WqyJh <781345688@qq.com>
Date: Mon, 20 Jan 2025 18:58:13 +0800
Subject: [PATCH 6/6] feat: support new ratio for all ui
---
relay/adaptor/openai/token.go | 8 +--
web/air/src/components/OperationSetting.js | 46 ++++++----------
.../Setting/component/OperationSetting.js | 52 +++++++------------
.../src/components/OperationSetting.js | 40 +-------------
4 files changed, 40 insertions(+), 106 deletions(-)
diff --git a/relay/adaptor/openai/token.go b/relay/adaptor/openai/token.go
index 7c8468b9..bda08577 100644
--- a/relay/adaptor/openai/token.go
+++ b/relay/adaptor/openai/token.go
@@ -3,14 +3,14 @@ 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
@@ -32,7 +32,7 @@ func InitTokenEncoders() {
if err != nil {
logger.FatalLog(fmt.Sprintf("failed to get gpt-4 token encoder: %s", err.Error()))
}
- for model := range billingratio.ModelRatio {
+ for model := range RatioMap {
if strings.HasPrefix(model, "gpt-3.5") {
tokenEncoderMap[model] = gpt35TokenEncoder
} else if strings.HasPrefix(model, "gpt-4o") {
diff --git a/web/air/src/components/OperationSetting.js b/web/air/src/components/OperationSetting.js
index 6356ac66..d13c882c 100644
--- a/web/air/src/components/OperationSetting.js
+++ b/web/air/src/components/OperationSetting.js
@@ -2,6 +2,13 @@ import React, { useEffect, useState } from 'react';
import { Divider, Form, Grid, Header } from 'semantic-ui-react';
import { API, showError, showSuccess, timestamp2string, verifyJSON } from '../helpers';
+const RATIO_MAPPING_EXAMPLE = {
+ 'gpt-4o-mini': {'input': 0.075, 'output': 0.3},
+ 'llama3-8b-8192(33)': {'input': 0.15, 'output': 0.3},
+ 'llama3-70b-8192(33)': {'input': 1.325, 'output': 1.749},
+};
+
+
const OperationSetting = () => {
let now = new Date();
let [inputs, setInputs] = useState({
@@ -10,8 +17,7 @@ const OperationSetting = () => {
QuotaForInvitee: 0,
QuotaRemindThreshold: 0,
PreConsumedQuota: 0,
- ModelRatio: '',
- CompletionRatio: '',
+ Ratio: '',
GroupRatio: '',
TopUpLink: '',
ChatLink: '',
@@ -35,7 +41,7 @@ const OperationSetting = () => {
if (success) {
let newInputs = {};
data.forEach((item) => {
- if (item.key === 'ModelRatio' || item.key === 'GroupRatio' || item.key === 'CompletionRatio') {
+ if (item.key === 'GroupRatio' || item.key === 'Ratio') {
item.value = JSON.stringify(JSON.parse(item.value), null, 2);
}
if (item.value === '{}') {
@@ -91,12 +97,12 @@ const OperationSetting = () => {
}
break;
case 'ratio':
- if (originInputs['ModelRatio'] !== inputs.ModelRatio) {
- if (!verifyJSON(inputs.ModelRatio)) {
- showError('模型倍率不是合法的 JSON 字符串');
+ if (originInputs['Ratio'] !== inputs.Ratio) {
+ if (!verifyJSON(inputs.Ratio)) {
+ showError('自定义倍率不是合法的 JSON 字符串');
return;
}
- await updateOption('ModelRatio', inputs.ModelRatio);
+ await updateOption('Ratio', inputs.Ratio);
}
if (originInputs['GroupRatio'] !== inputs.GroupRatio) {
if (!verifyJSON(inputs.GroupRatio)) {
@@ -105,13 +111,6 @@ const OperationSetting = () => {
}
await updateOption('GroupRatio', inputs.GroupRatio);
}
- if (originInputs['CompletionRatio'] !== inputs.CompletionRatio) {
- if (!verifyJSON(inputs.CompletionRatio)) {
- showError('补全倍率不是合法的 JSON 字符串');
- return;
- }
- await updateOption('CompletionRatio', inputs.CompletionRatio);
- }
break;
case 'quota':
if (originInputs['QuotaForNewUser'] !== inputs.QuotaForNewUser) {
@@ -346,24 +345,13 @@ const OperationSetting = () => {
-
-
-
diff --git a/web/berry/src/views/Setting/component/OperationSetting.js b/web/berry/src/views/Setting/component/OperationSetting.js
index 2bed715b..ee609d77 100644
--- a/web/berry/src/views/Setting/component/OperationSetting.js
+++ b/web/berry/src/views/Setting/component/OperationSetting.js
@@ -18,6 +18,12 @@ import { DateTimePicker } from "@mui/x-date-pickers/DateTimePicker";
import dayjs from "dayjs";
require("dayjs/locale/zh-cn");
+const RATIO_MAPPING_EXAMPLE = {
+ 'gpt-4o-mini': {'input': 0.075, 'output': 0.3},
+ 'llama3-8b-8192(33)': {'input': 0.15, 'output': 0.3},
+ 'llama3-70b-8192(33)': {'input': 1.325, 'output': 1.749},
+};
+
const OperationSetting = () => {
let now = new Date();
let [inputs, setInputs] = useState({
@@ -26,8 +32,7 @@ const OperationSetting = () => {
QuotaForInvitee: 0,
QuotaRemindThreshold: 0,
PreConsumedQuota: 0,
- ModelRatio: "",
- CompletionRatio: "",
+ Ratio: "",
GroupRatio: "",
TopUpLink: "",
ChatLink: "",
@@ -53,7 +58,7 @@ const OperationSetting = () => {
if (success) {
let newInputs = {};
data.forEach((item) => {
- if (item.key === "ModelRatio" || item.key === "GroupRatio" || item.key === "CompletionRatio") {
+ if (item.key === "GroupRatio" || item.key === "Ratio") {
item.value = JSON.stringify(JSON.parse(item.value), null, 2);
}
if (item.value === '{}') {
@@ -123,12 +128,12 @@ const OperationSetting = () => {
}
break;
case "ratio":
- if (originInputs["ModelRatio"] !== inputs.ModelRatio) {
- if (!verifyJSON(inputs.ModelRatio)) {
- showError("模型倍率不是合法的 JSON 字符串");
+ if (originInputs["Ratio"] !== inputs.Ratio) {
+ if (!verifyJSON(inputs.Ratio)) {
+ showError("自定义倍率不是合法的 JSON 字符串");
return;
}
- await updateOption("ModelRatio", inputs.ModelRatio);
+ await updateOption("Ratio", inputs.Ratio);
}
if (originInputs["GroupRatio"] !== inputs.GroupRatio) {
if (!verifyJSON(inputs.GroupRatio)) {
@@ -137,13 +142,6 @@ const OperationSetting = () => {
}
await updateOption("GroupRatio", inputs.GroupRatio);
}
- if (originInputs['CompletionRatio'] !== inputs.CompletionRatio) {
- if (!verifyJSON(inputs.CompletionRatio)) {
- showError('补全倍率不是合法的 JSON 字符串');
- return;
- }
- await updateOption('CompletionRatio', inputs.CompletionRatio);
- }
break;
case "quota":
if (originInputs["QuotaForNewUser"] !== inputs.QuotaForNewUser) {
@@ -501,28 +499,14 @@ const OperationSetting = () => {
-
-
-
diff --git a/web/default/src/components/OperationSetting.js b/web/default/src/components/OperationSetting.js
index 88cbf2d0..eaf7a234 100644
--- a/web/default/src/components/OperationSetting.js
+++ b/web/default/src/components/OperationSetting.js
@@ -16,8 +16,6 @@ const OperationSetting = () => {
QuotaForInvitee: 0,
QuotaRemindThreshold: 0,
PreConsumedQuota: 0,
- ModelRatio: '', // Deprecated
- CompletionRatio: '', // Deprecated
GroupRatio: '',
Ratio: '',
TopUpLink: '',
@@ -42,7 +40,7 @@ const OperationSetting = () => {
if (success) {
let newInputs = {};
data.forEach((item) => {
- if (item.key === 'ModelRatio' || item.key === 'GroupRatio' || item.key === 'CompletionRatio' || item.key === 'Ratio') {
+ if (item.key === 'GroupRatio' || item.key === 'Ratio') {
item.value = JSON.stringify(JSON.parse(item.value), null, 2);
}
if (item.value === '{}') {
@@ -98,13 +96,6 @@ const OperationSetting = () => {
}
break;
case 'ratio':
- if (originInputs['ModelRatio'] !== inputs.ModelRatio) {
- if (!verifyJSON(inputs.ModelRatio)) {
- showError('模型倍率不是合法的 JSON 字符串');
- return;
- }
- await updateOption('ModelRatio', inputs.ModelRatio);
- }
if (originInputs['GroupRatio'] !== inputs.GroupRatio) {
if (!verifyJSON(inputs.GroupRatio)) {
showError('分组倍率不是合法的 JSON 字符串');
@@ -112,13 +103,6 @@ const OperationSetting = () => {
}
await updateOption('GroupRatio', inputs.GroupRatio);
}
- if (originInputs['CompletionRatio'] !== inputs.CompletionRatio) {
- if (!verifyJSON(inputs.CompletionRatio)) {
- showError('补全倍率不是合法的 JSON 字符串');
- return;
- }
- await updateOption('CompletionRatio', inputs.CompletionRatio);
- }
if (originInputs['Ratio'] !== inputs.Ratio) {
if (!verifyJSON(inputs.Ratio)) {
showError('倍率不是合法的 JSON 字符串');
@@ -369,28 +353,6 @@ const OperationSetting = () => {
placeholder={`为一个 JSON 文本，键为模型名称，值为倍率结构，例如：\n${JSON.stringify(RATIO_MAPPING_EXAMPLE, null, 2)}`}
/>
-
-
-
-
-
-