refactor: replace log package with log/slog for structured logging and improve error handling

fix: trim suffix p for jimeng image model

LOGGING.md

@@ -0,0 +1,260 @@
+# 日志系统说明
+
+本项目使用 Go 标准库的 `log/slog` 实现结构化日志记录。
+
+## 📋 功能特性
+
+### 1. 标准的文件存储结构
+
+- **当前日志文件**: `oneapi.log` - 实时写入的日志文件
+- **归档日志文件**: `oneapi.2024-01-02-153045.log` - 自动轮转后的历史日志
+
+### 2. 自动日志轮转
+
+日志文件会在以下情况自动轮转：
+
+- **按大小轮转**: 当日志文件超过指定大小时（默认 100MB）
+- **启动时日期检查**: 程序启动时如果检测到日志文件是旧日期创建的，会自动轮转
+- **自动清理**: 只保留最近 N 个日志文件（默认 7 个）
+
+### 3. 结构化日志
+
+所有日志都包含以下结构化字段：
+
+```
+time=2024-01-02T15:30:45 level=INFO msg="user logged in" request_id=abc123 user_id=1001
+```
+
+### 4. 多种输出格式
+
+- **Text 格式** (默认): 人类可读的文本格式
+- **JSON 格式**: 便于日志分析工具解析
+
+### 5. 灵活的日志级别
+
+支持四个日志级别：
+- `DEBUG`: 调试信息
+- `INFO`: 一般信息
+- `WARN`: 警告信息
+- `ERROR`: 错误信息
+
+## ⚙️ 配置方式
+
+### 环境变量配置
+
+```bash
+# 日志目录（必需，否则只输出到控制台）
+--log-dir=./logs
+
+# 日志级别（可选，默认: INFO，DEBUG 模式除外）
+export LOG_LEVEL=DEBUG        # 可选值: DEBUG, INFO, WARN, ERROR
+
+# 日志格式（可选，默认: text）
+export LOG_FORMAT=json        # 可选值: text, json
+
+# 单个日志文件最大大小（可选，默认: 100，单位: MB）
+export LOG_MAX_SIZE_MB=200
+
+# 保留的日志文件数量（可选，默认: 7）
+export LOG_MAX_FILES=14
+
+# 启用调试模式（会自动将日志级别设为 DEBUG）
+export DEBUG=true
+```
+
+### 命令行参数
+
+```bash
+# 启动时指定日志目录
+./new-api --log-dir=./logs
+
+# 如果不指定日志目录，日志只输出到控制台
+./new-api
+```
+
+## 📝 使用示例
+
+### 基础使用
+
+```go
+import (
+    "context"
+    "github.com/QuantumNous/new-api/logger"
+)
+
+// 记录信息日志
+logger.LogInfo(ctx, "user registered successfully")
+
+// 记录警告日志
+logger.LogWarn(ctx, "API rate limit approaching")
+
+// 记录错误日志
+logger.LogError(ctx, "failed to connect to database")
+
+// 记录调试日志（只在 DEBUG 模式下输出）
+logger.LogDebug(ctx, "processing request with params: %v", params)
+
+// 记录系统日志（无 context）
+logger.LogSystemInfo("application started")
+logger.LogSystemError("critical system error")
+```
+
+### 日志输出示例
+
+**Text 格式** (易读格式):
+```
+[INFO] 2024/01/02 - 15:30:45 | SYSTEM | application started
+[INFO] 2024/01/02 - 15:30:46 | abc123 | user registered successfully
+[WARN] 2024/01/02 - 15:30:47 | def456 | API rate limit approaching | remaining=10, limit=100
+[ERROR] 2024/01/02 - 15:30:48 | ghi789 | failed to connect to database | error="connection timeout"
+```
+
+格式说明：`[级别] 时间 | 请求ID/组件 | 消息 | 额外属性（如有）`
+
+**JSON 格式**:
+```json
+{"time":"2024-01-02 15:30:45","level":"INFO","msg":"application started","request_id":"SYSTEM"}
+{"time":"2024-01-02 15:30:46","level":"INFO","msg":"user registered successfully","request_id":"abc123"}
+{"time":"2024-01-02 15:30:47","level":"WARN","msg":"API rate limit approaching","request_id":"def456"}
+```
+
+## 📂 日志文件结构
+
+```
+logs/
+├── oneapi.log                      # 当前活动日志文件
+├── oneapi.2024-01-01-090000.log  # 昨天的日志
+├── oneapi.2024-01-01-150000.log  # 昨天下午的日志（如果超过大小限制）
+├── oneapi.2023-12-31-090000.log  # 更早的日志
+└── ...                            # 最多保留配置数量的历史文件
+```
+
+## 🔄 日志轮转机制
+
+### 轮转触发条件
+
+1. **文件大小检查**: 每写入 1000 条日志后检查一次文件大小
+2. **启动时日期检查**: 程序启动时检查日志文件的修改日期，如果不是今天则轮转
+3. **自动清理**: 轮转时自动删除超过保留数量的旧日志文件
+
+> **注意**: 日志不会在运行时动态检查日期变化。如果需要每天自动轮转日志，建议：
+> - 使用定时任务（如 cron）每天重启服务
+> - 或者配置较小的日志文件大小，让它自动按大小轮转
+
+### 轮转流程
+
+1. 检测到需要轮转时，关闭当前日志文件
+2. 将 `oneapi.log` 重命名为 `oneapi.YYYY-MM-DD-HHmmss.log`
+3. 创建新的 `oneapi.log` 文件
+4. 异步清理超过数量限制的旧日志文件
+5. 记录轮转事件到新日志文件
+
+## 🎯 最佳实践
+
+### 1. 生产环境配置
+
+```bash
+# 使用 INFO 级别，避免过多调试信息
+export LOG_LEVEL=INFO
+
+# 使用 JSON 格式，便于日志分析工具处理
+export LOG_FORMAT=json
+
+# 设置合适的文件大小和保留数量
+export LOG_MAX_SIZE_MB=500
+export LOG_MAX_FILES=30
+
+# 指定日志目录
+./new-api --log-dir=/var/log/oneapi
+```
+
+### 2. 开发环境配置
+
+```bash
+# 使用 DEBUG 级别查看详细信息
+export DEBUG=true
+
+# 使用 Text 格式，便于阅读
+export LOG_FORMAT=text
+
+# 较小的文件大小和保留数量
+export LOG_MAX_SIZE_MB=50
+export LOG_MAX_FILES=7
+
+./new-api --log-dir=./logs
+```
+
+### 3. 容器环境配置
+
+```bash
+# 只输出到标准输出，由容器运行时管理日志
+./new-api
+
+# 或者使用 JSON 格式便于日志收集系统处理
+export LOG_FORMAT=json
+./new-api
+```
+
+## 🔍 日志分析
+
+### 使用 grep 分析文本日志
+
+```bash
+# 查找错误日志
+grep '\[ERROR\]' logs/oneapi.log
+
+# 查找特定请求的所有日志
+grep 'abc123' logs/*.log
+
+# 查看最近的警告和错误
+tail -f logs/oneapi.log | grep -E '\[(WARN|ERROR)\]'
+
+# 查找包含特定关键词的日志
+grep 'database' logs/oneapi.log
+
+# 查看今天的所有错误
+grep "\[ERROR\] $(date +%Y/%m/%d)" logs/oneapi.log
+```
+
+### 使用 jq 分析 JSON 日志
+
+```bash
+# 提取所有错误日志
+cat logs/oneapi.log | jq 'select(.level=="ERROR")'
+
+# 统计各级别日志数量
+cat logs/oneapi.log | jq -r '.level' | sort | uniq -c
+
+# 查找特定时间范围的日志
+cat logs/oneapi.log | jq 'select(.time >= "2024-01-02 15:00:00" and .time <= "2024-01-02 16:00:00")'
+```
+
+## 📊 性能优化
+
+1. **异步日志轮转**: 轮转操作在后台 goroutine 中执行，不阻塞主程序
+2. **批量写入检查**: 每 1000 次写入才检查一次轮转条件，减少 I/O 开销
+3. **读写锁**: 使用 `sync.RWMutex` 保护日志器，提高并发性能
+4. **零分配**: `slog` 库在大多数情况下实现零内存分配
+
+## 🚨 故障排查
+
+### 日志文件未创建
+
+- 检查日志目录是否存在且有写入权限
+- 确认启动时指定了 `--log-dir` 参数
+
+### 日志文件过多
+
+- 调整 `LOG_MAX_FILES` 环境变量
+- 手动清理不需要的旧日志文件
+
+### 日志级别不正确
+
+- 检查 `LOG_LEVEL` 环境变量是否正确设置
+- 确认 `DEBUG` 环境变量的值（会覆盖 LOG_LEVEL）
+
+## 📖 相关文档
+
+- [Go slog 官方文档](https://pkg.go.dev/log/slog)
+- [结构化日志最佳实践](https://go.dev/blog/slog)
+

README.md

@@ -141,6 +141,7 @@ New API提供了丰富的功能，详细特性请参考[特性说明](https://do
 - `NOTIFICATION_LIMIT_DURATION_MINUTE`：邮件等通知限制持续时间，默认 `10`分钟
 - `NOTIFY_LIMIT_COUNT`：用户通知在指定持续时间内的最大数量，默认 `2`
 - `ERROR_LOG_ENABLED=true`: 是否记录并显示错误日志，默认`false`
+- `TASK_PRICE_PATCH=sora-2-all,sora-2-pro-all`: 异步任务设置某些模型按次计费，多个模型用逗号分隔，例如`sora-2-all,sora-2-pro-all`，表示sora-2-all和sora-2-pro-all模型异步任务仅按次计费，不按秒等计费。
 
 ## 部署
 

common/constants.go

@@ -159,14 +159,15 @@ var (
 	GlobalWebRateLimitNum      int
 	GlobalWebRateLimitDuration int64
 
+	CriticalRateLimitEnable   bool
+	CriticalRateLimitNum            = 20
+	CriticalRateLimitDuration int64 = 20 * 60
+
 	UploadRateLimitNum            = 10
 	UploadRateLimitDuration int64 = 60
 
 	DownloadRateLimitNum            = 10
 	DownloadRateLimitDuration int64 = 60
-
-	CriticalRateLimitNum            = 20
-	CriticalRateLimitDuration int64 = 20 * 60
 )
 
 var RateLimitKeyExpirationDuration = 20 * time.Minute

common/init.go

@@ -3,7 +3,7 @@ package common
 import (
 	"flag"
 	"fmt"
-	"log"
+	"log/slog"
 	"os"
 	"path/filepath"
 	"strconv"
@@ -43,9 +43,10 @@ func InitEnv() {
 	if os.Getenv("SESSION_SECRET") != "" {
 		ss := os.Getenv("SESSION_SECRET")
 		if ss == "random_string" {
-			log.Println("WARNING: SESSION_SECRET is set to the default value 'random_string', please change it to a random string.")
+			slog.Warn("SESSION_SECRET is set to the default value 'random_string', please change it to a random string.")
-			log.Println("警告：SESSION_SECRET被设置为默认值'random_string'，请修改为随机字符串。")
+			slog.Warn("警告：SESSION_SECRET被设置为默认值'random_string'，请修改为随机字符串。")
-			log.Fatal("Please set SESSION_SECRET to a random string.")
+			slog.Error("Please set SESSION_SECRET to a random string.")
+			os.Exit(1)
 		} else {
 			SessionSecret = ss
 		}
@@ -62,12 +63,14 @@ func InitEnv() {
 		var err error
 		*LogDir, err = filepath.Abs(*LogDir)
 		if err != nil {
-			log.Fatal(err)
+			slog.Error("failed to get absolute path for log directory", "error", err)
+			os.Exit(1)
 		}
 		if _, err := os.Stat(*LogDir); os.IsNotExist(err) {
 			err = os.Mkdir(*LogDir, 0777)
 			if err != nil {
-				log.Fatal(err)
+				slog.Error("failed to create log directory", "error", err)
+				os.Exit(1)
 			}
 		}
 	}
@@ -99,6 +102,9 @@ func InitEnv() {
 	GlobalWebRateLimitNum = GetEnvOrDefault("GLOBAL_WEB_RATE_LIMIT", 60)
 	GlobalWebRateLimitDuration = int64(GetEnvOrDefault("GLOBAL_WEB_RATE_LIMIT_DURATION", 180))
 
+	CriticalRateLimitEnable = GetEnvOrDefaultBool("CRITICAL_RATE_LIMIT_ENABLE", true)
+	CriticalRateLimitNum = GetEnvOrDefault("CRITICAL_RATE_LIMIT", 20)
+	CriticalRateLimitDuration = int64(GetEnvOrDefault("CRITICAL_RATE_LIMIT_DURATION", 20*60))
 	initConstantEnv()
 }
 

common/sys_log.go

@@ -2,6 +2,7 @@ package common
 
 import (
 	"fmt"
+	"log/slog"
 	"os"
 	"time"
 
@@ -9,18 +10,16 @@ import (
 )
 
 func SysLog(s string) {
-	t := time.Now()
+	slog.Info(s, "component", "system")
-	_, _ = fmt.Fprintf(gin.DefaultWriter, "[SYS] %v | %s \n", t.Format("2006/01/02 - 15:04:05"), s)
 }
 
 func SysError(s string) {
-	t := time.Now()
+	slog.Error(s, "component", "system")
-	_, _ = fmt.Fprintf(gin.DefaultErrorWriter, "[SYS] %v | %s \n", t.Format("2006/01/02 - 15:04:05"), s)
 }
 
 func FatalLog(v ...any) {
-	t := time.Now()
+	msg := fmt.Sprint(v...)
-	_, _ = fmt.Fprintf(gin.DefaultErrorWriter, "[FATAL] %v | %v \n", t.Format("2006/01/02 - 15:04:05"), v)
+	slog.Error(msg, "component", "system", "level", "fatal")
 	os.Exit(1)
 }
 

controller/channel-test.go

@@ -617,6 +617,10 @@ func TestAllChannels(c *gin.Context) {
 var autoTestChannelsOnce sync.Once
 
 func AutomaticallyTestChannels() {
+	// 只在Master节点定时测试渠道
+	if !common.IsMasterNode {
+		return
+	}
 	autoTestChannelsOnce.Do(func() {
 		for {
 			if !operation_setting.GetMonitorSetting().AutoTestChannelEnabled {

controller/channel.go

@@ -649,13 +649,15 @@ func DeleteDisabledChannel(c *gin.Context) {
 }
 
 type ChannelTag struct {
-	Tag          string  `json:"tag"`
+	Tag            string  `json:"tag"`
-	NewTag       *string `json:"new_tag"`
+	NewTag         *string `json:"new_tag"`
-	Priority     *int64  `json:"priority"`
+	Priority       *int64  `json:"priority"`
-	Weight       *uint   `json:"weight"`
+	Weight         *uint   `json:"weight"`
-	ModelMapping *string `json:"model_mapping"`
+	ModelMapping   *string `json:"model_mapping"`
-	Models       *string `json:"models"`
+	Models         *string `json:"models"`
-	Groups       *string `json:"groups"`
+	Groups         *string `json:"groups"`
+	ParamOverride  *string `json:"param_override"`
+	HeaderOverride *string `json:"header_override"`
 }
 
 func DisableTagChannels(c *gin.Context) {
@@ -721,7 +723,29 @@ func EditTagChannels(c *gin.Context) {
 		})
 		return
 	}
-	err = model.EditChannelByTag(channelTag.Tag, channelTag.NewTag, channelTag.ModelMapping, channelTag.Models, channelTag.Groups, channelTag.Priority, channelTag.Weight)
+	if channelTag.ParamOverride != nil {
+		trimmed := strings.TrimSpace(*channelTag.ParamOverride)
+		if trimmed != "" && !json.Valid([]byte(trimmed)) {
+			c.JSON(http.StatusOK, gin.H{
+				"success": false,
+				"message": "参数覆盖必须是合法的 JSON 格式",
+			})
+			return
+		}
+		channelTag.ParamOverride = common.GetPointer[string](trimmed)
+	}
+	if channelTag.HeaderOverride != nil {
+		trimmed := strings.TrimSpace(*channelTag.HeaderOverride)
+		if trimmed != "" && !json.Valid([]byte(trimmed)) {
+			c.JSON(http.StatusOK, gin.H{
+				"success": false,
+				"message": "请求头覆盖必须是合法的 JSON 格式",
+			})
+			return
+		}
+		channelTag.HeaderOverride = common.GetPointer[string](trimmed)
+	}
+	err = model.EditChannelByTag(channelTag.Tag, channelTag.NewTag, channelTag.ModelMapping, channelTag.Models, channelTag.Groups, channelTag.Priority, channelTag.Weight, channelTag.ParamOverride, channelTag.HeaderOverride)
 	if err != nil {
 		common.ApiError(c, err)
 		return

dto/claude.go

@@ -510,11 +510,44 @@ func (c *ClaudeResponse) GetClaudeError() *types.ClaudeError {
 }
 
 type ClaudeUsage struct {
-	InputTokens              int                  `json:"input_tokens"`
+	InputTokens              int                       `json:"input_tokens"`
-	CacheCreationInputTokens int                  `json:"cache_creation_input_tokens"`
+	CacheCreationInputTokens int                       `json:"cache_creation_input_tokens"`
-	CacheReadInputTokens     int                  `json:"cache_read_input_tokens"`
+	CacheReadInputTokens     int                       `json:"cache_read_input_tokens"`
-	OutputTokens             int                  `json:"output_tokens"`
+	OutputTokens             int                       `json:"output_tokens"`
-	ServerToolUse            *ClaudeServerToolUse `json:"server_tool_use,omitempty"`
+	CacheCreation            *ClaudeCacheCreationUsage `json:"cache_creation,omitempty"`
+	// claude cache 1h
+	ClaudeCacheCreation5mTokens int                  `json:"claude_cache_creation_5_m_tokens"`
+	ClaudeCacheCreation1hTokens int                  `json:"claude_cache_creation_1_h_tokens"`
+	ServerToolUse               *ClaudeServerToolUse `json:"server_tool_use,omitempty"`
+}
+
+type ClaudeCacheCreationUsage struct {
+	Ephemeral5mInputTokens int `json:"ephemeral_5m_input_tokens,omitempty"`
+	Ephemeral1hInputTokens int `json:"ephemeral_1h_input_tokens,omitempty"`
+}
+
+func (u *ClaudeUsage) GetCacheCreation5mTokens() int {
+	if u == nil || u.CacheCreation == nil {
+		return 0
+	}
+	return u.CacheCreation.Ephemeral5mInputTokens
+}
+
+func (u *ClaudeUsage) GetCacheCreation1hTokens() int {
+	if u == nil || u.CacheCreation == nil {
+		return 0
+	}
+	return u.CacheCreation.Ephemeral1hInputTokens
+}
+
+func (u *ClaudeUsage) GetCacheCreationTotalTokens() int {
+	if u == nil {
+		return 0
+	}
+	if u.CacheCreationInputTokens > 0 {
+		return u.CacheCreationInputTokens
+	}
+	return u.GetCacheCreation5mTokens() + u.GetCacheCreation1hTokens()
 }
 
 type ClaudeServerToolUse struct {

dto/openai_request.go

@@ -232,10 +232,13 @@ func (r *GeneralOpenAIRequest) GetSystemRoleName() string {
 	return "system"
 }
 
+const CustomType = "custom"
+
 type ToolCallRequest struct {
 	ID       string          `json:"id,omitempty"`
 	Type     string          `json:"type"`
-	Function FunctionRequest `json:"function"`
+	Function FunctionRequest `json:"function,omitempty"`
+	Custom   json.RawMessage `json:"custom,omitempty"`
 }
 
 type FunctionRequest struct {

dto/openai_response.go

@@ -230,6 +230,11 @@ type Usage struct {
 	InputTokens            int                `json:"input_tokens"`
 	OutputTokens           int                `json:"output_tokens"`
 	InputTokensDetails     *InputTokenDetails `json:"input_tokens_details"`
+
+	// claude cache 1h
+	ClaudeCacheCreation5mTokens int `json:"claude_cache_creation_5_m_tokens"`
+	ClaudeCacheCreation1hTokens int `json:"claude_cache_creation_1_h_tokens"`
+
 	// OpenRouter Params
 	Cost any `json:"cost,omitempty"`
 }

logger/logger.go

@@ -5,9 +5,11 @@ import (
 	"encoding/json"
 	"fmt"
 	"io"
-	"log"
+	"log/slog"
 	"os"
 	"path/filepath"
+	"sort"
+	"strings"
 	"sync"
 	"time"
 
@@ -19,79 +21,561 @@ import (
 )
 
 const (
-	loggerINFO  = "INFO"
+	// 日志轮转配置
-	loggerWarn  = "WARN"
+	defaultMaxLogSize   = 100 * 1024 * 1024 // 100MB
-	loggerError = "ERR"
+	defaultMaxLogFiles  = 7                 // 保留最近7个日志文件
-	loggerDebug = "DEBUG"
+	defaultLogFileName  = "newapi.log"
+	checkRotateInterval = 1000 // 每1000次写入检查一次是否需要轮转
 )
 
-const maxLogCount = 1000000
+var (
+	logMutex        sync.RWMutex
+	rotateCheckLock sync.Mutex
+	defaultLogger   *slog.Logger
+	logFile         *os.File
+	logFilePath     string
+	logDirPath      string
+	writeCount      int64
+	maxLogSize      int64 = defaultMaxLogSize
+	maxLogFiles     int   = defaultMaxLogFiles
+	useJSONFormat   bool
+)
 
-var logCount int
+func init() {
-var setupLogLock sync.Mutex
+	// Initialize with a text handler to stdout
-var setupLogWorking bool
+	handler := createHandler(os.Stdout)
+	defaultLogger = slog.New(handler)
+	slog.SetDefault(defaultLogger)
+}
 
+// SetupLogger 初始化日志系统
 func SetupLogger() {
-	defer func() {
+	logMutex.Lock()
-		setupLogWorking = false
+	defer logMutex.Unlock()
-	}()
+
-	if *common.LogDir != "" {
+	// 读取环境变量配置
-		ok := setupLogLock.TryLock()
+	if maxSize := os.Getenv("LOG_MAX_SIZE_MB"); maxSize != "" {
-		if !ok {
+		if size, err := fmt.Sscanf(maxSize, "%d", &maxLogSize); err == nil && size > 0 {
-			log.Println("setup log is already working")
+			maxLogSize = maxLogSize * 1024 * 1024 // 转换为字节
-			return
 		}
-		defer func() {
+	}
-			setupLogLock.Unlock()
+	if maxFiles := os.Getenv("LOG_MAX_FILES"); maxFiles != "" {
-		}()
+		fmt.Sscanf(maxFiles, "%d", &maxLogFiles)
-		logPath := filepath.Join(*common.LogDir, fmt.Sprintf("oneapi-%s.log", time.Now().Format("20060102150405")))
+	}
-		fd, err := os.OpenFile(logPath, os.O_APPEND|os.O_CREATE|os.O_WRONLY, 0644)
+	if os.Getenv("LOG_FORMAT") == "json" {
-		if err != nil {
+		useJSONFormat = true
-			log.Fatal("failed to open log file")
+	}
+
+	if *common.LogDir == "" {
+		// 如果没有配置日志目录，只输出到标准输出
+		handler := createHandler(os.Stdout)
+		defaultLogger = slog.New(handler)
+		slog.SetDefault(defaultLogger)
+		return
+	}
+
+	logDirPath = *common.LogDir
+	logFilePath = filepath.Join(logDirPath, defaultLogFileName)
+
+	// 检查日志文件是否需要按日期轮转（仅在启动时检查）
+	if err := checkAndRotateOnStartup(); err != nil {
+		slog.Error("failed to check log file on startup", "error", err)
+	}
+
+	// 打开或创建日志文件
+	if err := openLogFile(); err != nil {
+		slog.Error("failed to open log file", "error", err)
+		return
+	}
+
+	// 创建多路输出（控制台 + 文件）
+	multiWriter := io.MultiWriter(os.Stdout, logFile)
+
+	// 更新 gin 的默认输出
+	gin.DefaultWriter = multiWriter
+	gin.DefaultErrorWriter = multiWriter
+
+	// 更新 slog handler
+	handler := createHandler(multiWriter)
+	defaultLogger = slog.New(handler)
+	slog.SetDefault(defaultLogger)
+
+	slog.Info("logger initialized",
+		"log_dir", logDirPath,
+		"max_size_mb", maxLogSize/(1024*1024),
+		"max_files", maxLogFiles,
+		"format", getLogFormat())
+}
+
+// createHandler 创建日志处理器
+func createHandler(w io.Writer) slog.Handler {
+	if useJSONFormat {
+		opts := &slog.HandlerOptions{
+			Level: getLogLevel(),
 		}
-		gin.DefaultWriter = io.MultiWriter(os.Stdout, fd)
+		return slog.NewJSONHandler(w, opts)
-		gin.DefaultErrorWriter = io.MultiWriter(os.Stderr, fd)
+	}
+	return NewReadableTextHandler(w, getLogLevel())
+}
+
+// ReadableTextHandler 自定义的易读文本处理器
+type ReadableTextHandler struct {
+	w     io.Writer
+	level slog.Level
+	mu    sync.Mutex
+}
+
+// NewReadableTextHandler 创建一个新的易读文本处理器
+func NewReadableTextHandler(w io.Writer, level slog.Level) *ReadableTextHandler {
+	return &ReadableTextHandler{
+		w:     w,
+		level: level,
 	}
 }
 
-func LogInfo(ctx context.Context, msg string) {
+// Enabled 检查是否启用该级别
-	logHelper(ctx, loggerINFO, msg)
+func (h *ReadableTextHandler) Enabled(_ context.Context, level slog.Level) bool {
+	return level >= h.level
 }
 
-func LogWarn(ctx context.Context, msg string) {
+// Handle 处理日志记录
-	logHelper(ctx, loggerWarn, msg)
+func (h *ReadableTextHandler) Handle(_ context.Context, r slog.Record) error {
+	h.mu.Lock()
+	defer h.mu.Unlock()
+
+	// 格式: [LEVEL] YYYY/MM/DD - HH:mm:ss | request_id | message | key=value ...
+	buf := make([]byte, 0, 256)
+
+	// 日志级别
+	level := r.Level.String()
+	switch r.Level {
+	case slog.LevelDebug:
+		level = "DEBUG"
+	case slog.LevelInfo:
+		level = "INFO"
+	case slog.LevelWarn:
+		level = "WARN"
+	case slog.LevelError:
+		level = "ERROR"
+	}
+	buf = append(buf, '[')
+	buf = append(buf, level...)
+	buf = append(buf, "] "...)
+
+	// 时间
+	buf = append(buf, r.Time.Format("2006/01/02 - 15:04:05")...)
+	buf = append(buf, " | "...)
+
+	// 提取 request_id 和 component
+	var requestID, component string
+	otherAttrs := make([]slog.Attr, 0)
+
+	r.Attrs(func(a slog.Attr) bool {
+		switch a.Key {
+		case "request_id":
+			requestID = a.Value.String()
+		case "component":
+			component = a.Value.String()
+		default:
+			otherAttrs = append(otherAttrs, a)
+		}
+		return true
+	})
+
+	// 输出 request_id 或 component
+	if requestID != "" {
+		buf = append(buf, requestID...)
+		buf = append(buf, " | "...)
+	} else if component != "" {
+		buf = append(buf, component...)
+		buf = append(buf, " | "...)
+	}
+
+	// 消息
+	buf = append(buf, r.Message...)
+
+	// 其他属性
+	if len(otherAttrs) > 0 {
+		buf = append(buf, " | "...)
+		for i, a := range otherAttrs {
+			if i > 0 {
+				buf = append(buf, ", "...)
+			}
+			buf = append(buf, a.Key...)
+			buf = append(buf, '=')
+			buf = appendValue(buf, a.Value)
+		}
+	}
+
+	buf = append(buf, '\n')
+	_, err := h.w.Write(buf)
+	return err
 }
 
-func LogError(ctx context.Context, msg string) {
+// appendValue 追加值到缓冲区
-	logHelper(ctx, loggerError, msg)
+func appendValue(buf []byte, v slog.Value) []byte {
+	switch v.Kind() {
+	case slog.KindString:
+		s := v.String()
+		// 如果字符串包含空格或特殊字符，加引号
+		if strings.ContainsAny(s, " \t\n\r,=") {
+			buf = append(buf, '"')
+			buf = append(buf, s...)
+			buf = append(buf, '"')
+		} else {
+			buf = append(buf, s...)
+		}
+	case slog.KindInt64:
+		buf = append(buf, fmt.Sprintf("%d", v.Int64())...)
+	case slog.KindUint64:
+		buf = append(buf, fmt.Sprintf("%d", v.Uint64())...)
+	case slog.KindFloat64:
+		buf = append(buf, fmt.Sprintf("%g", v.Float64())...)
+	case slog.KindBool:
+		buf = append(buf, fmt.Sprintf("%t", v.Bool())...)
+	case slog.KindDuration:
+		buf = append(buf, v.Duration().String()...)
+	case slog.KindTime:
+		buf = append(buf, v.Time().Format("2006-01-02 15:04:05")...)
+	default:
+		buf = append(buf, fmt.Sprintf("%v", v.Any())...)
+	}
+	return buf
 }
 
-func LogDebug(ctx context.Context, msg string, args ...any) {
+// WithAttrs 返回一个新的处理器，包含指定的属性
-	msg = fmt.Sprintf(msg, args...)
+func (h *ReadableTextHandler) WithAttrs(attrs []slog.Attr) slog.Handler {
+	// 简化实现：不支持 With
+	return h
+}
+
+// WithGroup 返回一个新的处理器，使用指定的组
+func (h *ReadableTextHandler) WithGroup(name string) slog.Handler {
+	// 简化实现：不支持组
+	return h
+}
+
+// checkAndRotateOnStartup 启动时检查日志文件是否需要按日期轮转
+func checkAndRotateOnStartup() error {
+	// 检查日志文件是否存在
+	fileInfo, err := os.Stat(logFilePath)
+	if err != nil {
+		if os.IsNotExist(err) {
+			// 文件不存在，不需要轮转
+			return nil
+		}
+		return fmt.Errorf("failed to stat log file: %w", err)
+	}
+
+	// 获取文件的修改时间
+	modTime := fileInfo.ModTime()
+	modDate := modTime.Format("2006-01-02")
+	today := time.Now().Format("2006-01-02")
+
+	// 如果文件的日期和今天不同，进行轮转
+	if modDate != today {
+		// 生成归档文件名（使用文件的修改日期）
+		timestamp := modTime.Format("2006-01-02-150405")
+		archivePath := filepath.Join(logDirPath, fmt.Sprintf("newapi.%s.log", timestamp))
+
+		// 重命名日志文件
+		if err := os.Rename(logFilePath, archivePath); err != nil {
+			return fmt.Errorf("failed to archive old log file: %w", err)
+		}
+
+		slog.Info("rotated old log file on startup",
+			"archive", archivePath,
+			"reason", "date changed")
+
+		// 清理旧的日志文件
+		gopool.Go(func() {
+			cleanOldLogFiles()
+		})
+	}
+
+	return nil
+}
+
+// openLogFile 打开日志文件
+func openLogFile() error {
+	// 关闭旧的日志文件
+	if logFile != nil {
+		logFile.Close()
+	}
+
+	// 打开新的日志文件
+	fd, err := os.OpenFile(logFilePath, os.O_APPEND|os.O_CREATE|os.O_WRONLY, 0644)
+	if err != nil {
+		return fmt.Errorf("failed to open log file: %w", err)
+	}
+
+	logFile = fd
+	writeCount = 0
+	return nil
+}
+
+// rotateLogFile 轮转日志文件
+func rotateLogFile() error {
+	if logFile == nil {
+		return nil
+	}
+
+	rotateCheckLock.Lock()
+	defer rotateCheckLock.Unlock()
+
+	// 获取当前日志文件信息
+	fileInfo, err := logFile.Stat()
+	if err != nil {
+		return fmt.Errorf("failed to stat log file: %w", err)
+	}
+
+	// 检查文件大小是否需要轮转
+	if fileInfo.Size() < maxLogSize {
+		return nil
+	}
+
+	// 关闭当前日志文件
+	logFile.Close()
+
+	// 生成归档文件名
+	timestamp := time.Now().Format("2006-01-02-150405")
+	archivePath := filepath.Join(logDirPath, fmt.Sprintf("newapi.%s.log", timestamp))
+
+	// 重命名当前日志文件为归档文件
+	if err := os.Rename(logFilePath, archivePath); err != nil {
+		// 如果重命名失败，尝试复制
+		if copyErr := copyFile(logFilePath, archivePath); copyErr != nil {
+			return fmt.Errorf("failed to archive log file: %w", err)
+		}
+		os.Truncate(logFilePath, 0)
+	}
+
+	// 清理旧的日志文件
+	gopool.Go(func() {
+		cleanOldLogFiles()
+	})
+
+	// 打开新的日志文件
+	if err := openLogFile(); err != nil {
+		return err
+	}
+
+	// 重新设置日志输出
+	multiWriter := io.MultiWriter(os.Stdout, logFile)
+	gin.DefaultWriter = multiWriter
+	gin.DefaultErrorWriter = multiWriter
+
+	handler := createHandler(multiWriter)
+	logMutex.Lock()
+	defaultLogger = slog.New(handler)
+	slog.SetDefault(defaultLogger)
+	logMutex.Unlock()
+
+	slog.Info("log file rotated",
+		"reason", "size limit reached",
+		"archive", archivePath)
+
+	return nil
+}
+
+// cleanOldLogFiles 清理旧的日志文件
+func cleanOldLogFiles() {
+	if logDirPath == "" {
+		return
+	}
+
+	files, err := os.ReadDir(logDirPath)
+	if err != nil {
+		slog.Error("failed to read log directory", "error", err)
+		return
+	}
+
+	// 收集所有归档日志文件
+	var logFiles []os.DirEntry
+	for _, file := range files {
+		if !file.IsDir() && strings.HasPrefix(file.Name(), "newapi.") &&
+			strings.HasSuffix(file.Name(), ".log") &&
+			file.Name() != defaultLogFileName {
+			logFiles = append(logFiles, file)
+		}
+	}
+
+	// 如果归档文件数量超过限制，删除最旧的
+	if len(logFiles) > maxLogFiles {
+		// 按名称排序（文件名包含时间戳）
+		sort.Slice(logFiles, func(i, j int) bool {
+			return logFiles[i].Name() < logFiles[j].Name()
+		})
+
+		// 删除最旧的文件
+		deleteCount := len(logFiles) - maxLogFiles
+		for i := 0; i < deleteCount; i++ {
+			filePath := filepath.Join(logDirPath, logFiles[i].Name())
+			if err := os.Remove(filePath); err != nil {
+				slog.Error("failed to remove old log file",
+					"file", filePath,
+					"error", err)
+			} else {
+				slog.Info("removed old log file", "file", logFiles[i].Name())
+			}
+		}
+	}
+}
+
+// copyFile 复制文件
+func copyFile(src, dst string) error {
+	sourceFile, err := os.Open(src)
+	if err != nil {
+		return err
+	}
+	defer sourceFile.Close()
+
+	destFile, err := os.Create(dst)
+	if err != nil {
+		return err
+	}
+	defer destFile.Close()
+
+	_, err = io.Copy(destFile, sourceFile)
+	return err
+}
+
+// getLogLevel 获取日志级别
+func getLogLevel() slog.Level {
+	// 支持环境变量配置
+	if level := os.Getenv("LOG_LEVEL"); level != "" {
+		switch strings.ToUpper(level) {
+		case "DEBUG":
+			return slog.LevelDebug
+		case "INFO":
+			return slog.LevelInfo
+		case "WARN", "WARNING":
+			return slog.LevelWarn
+		case "ERROR":
+			return slog.LevelError
+		}
+	}
+
 	if common.DebugEnabled {
-		logHelper(ctx, loggerDebug, msg)
+		return slog.LevelDebug
+	}
+	return slog.LevelInfo
+}
+
+// getLogFormat 获取日志格式
+func getLogFormat() string {
+	if useJSONFormat {
+		return "json"
+	}
+	return "text"
+}
+
+// checkAndRotateLog 检查并轮转日志
+func checkAndRotateLog() {
+	if logFile == nil {
+		return
+	}
+
+	writeCount++
+	if writeCount%checkRotateInterval == 0 {
+		gopool.Go(func() {
+			if err := rotateLogFile(); err != nil {
+				slog.Error("failed to rotate log file", "error", err)
+			}
+		})
 	}
 }
 
-func logHelper(ctx context.Context, level string, msg string) {
+// LogInfo 记录信息级别日志
-	writer := gin.DefaultErrorWriter
+func LogInfo(ctx context.Context, msg string) {
-	if level == loggerINFO {
+	if ctx == nil {
-		writer = gin.DefaultWriter
+		ctx = context.Background()
+	}
+	id := getRequestID(ctx)
+	logMutex.RLock()
+	logger := defaultLogger
+	logMutex.RUnlock()
+	logger.InfoContext(ctx, msg, "request_id", id)
+	checkAndRotateLog()
+}
+
+// LogWarn 记录警告级别日志
+func LogWarn(ctx context.Context, msg string) {
+	if ctx == nil {
+		ctx = context.Background()
+	}
+	id := getRequestID(ctx)
+	logMutex.RLock()
+	logger := defaultLogger
+	logMutex.RUnlock()
+	logger.WarnContext(ctx, msg, "request_id", id)
+	checkAndRotateLog()
+}
+
+// LogError 记录错误级别日志
+func LogError(ctx context.Context, msg string) {
+	if ctx == nil {
+		ctx = context.Background()
+	}
+	id := getRequestID(ctx)
+	logMutex.RLock()
+	logger := defaultLogger
+	logMutex.RUnlock()
+	logger.ErrorContext(ctx, msg, "request_id", id)
+	checkAndRotateLog()
+}
+
+// LogSystemInfo 记录系统信息
+func LogSystemInfo(msg string) {
+	logMutex.RLock()
+	logger := defaultLogger
+	logMutex.RUnlock()
+	logger.Info(msg, "request_id", "SYSTEM")
+	checkAndRotateLog()
+}
+
+// LogSystemError 记录系统错误
+func LogSystemError(msg string) {
+	logMutex.RLock()
+	logger := defaultLogger
+	logMutex.RUnlock()
+	logger.Error(msg, "request_id", "SYSTEM")
+	checkAndRotateLog()
+}
+
+// LogDebug 记录调试级别日志
+func LogDebug(ctx context.Context, msg string, args ...any) {
+	if !common.DebugEnabled && getLogLevel() > slog.LevelDebug {
+		return
+	}
+
+	if ctx == nil {
+		ctx = context.Background()
+	}
+	id := getRequestID(ctx)
+	if len(args) > 0 {
+		msg = fmt.Sprintf(msg, args...)
+	}
+	logMutex.RLock()
+	logger := defaultLogger
+	logMutex.RUnlock()
+	logger.DebugContext(ctx, msg, "request_id", id)
+	checkAndRotateLog()
+}
+
+// getRequestID 从上下文中获取请求ID
+func getRequestID(ctx context.Context) string {
+	if ctx == nil {
+		return "SYSTEM"
 	}
 	id := ctx.Value(common.RequestIdKey)
 	if id == nil {
-		id = "SYSTEM"
+		return "SYSTEM"
 	}
-	now := time.Now()
+	if strID, ok := id.(string); ok {
-	_, _ = fmt.Fprintf(writer, "[%s] %v | %s | %s \n", level, now.Format("2006/01/02 - 15:04:05"), id, msg)
+		return strID
-	logCount++ // we don't need accurate count, so no lock here
-	if logCount > maxLogCount && !setupLogWorking {
-		logCount = 0
-		setupLogWorking = true
-		gopool.Go(func() {
-			SetupLogger()
-		})
 	}
+	return "SYSTEM"
 }
 
 func LogQuota(quota int) string {

main.go

@@ -4,7 +4,7 @@ import (
 	"bytes"
 	"embed"
 	"fmt"
-	"log"
+	"log/slog"
 	"net/http"
 	"os"
 	"strconv"
@@ -118,7 +118,9 @@ func main() {
 
 	if os.Getenv("ENABLE_PPROF") == "true" {
 		gopool.Go(func() {
-			log.Println(http.ListenAndServe("0.0.0.0:8005", nil))
+			if err := http.ListenAndServe("0.0.0.0:8005", nil); err != nil {
+				slog.Error("pprof server failed", "error", err)
+			}
 		})
 		go common.Monitor()
 		common.SysLog("pprof enabled")
@@ -127,7 +129,7 @@ func main() {
 	// Initialize HTTP server
 	server := gin.New()
 	server.Use(gin.CustomRecovery(func(c *gin.Context, err any) {
-		common.SysLog(fmt.Sprintf("panic detected: %v", err))
+		logger.LogSystemError(fmt.Sprintf("panic detected: %v", err))
 		c.JSON(http.StatusInternalServerError, gin.H{
 			"error": gin.H{
 				"message": fmt.Sprintf("Panic detected, error: %v. Please submit a issue here: https://github.com/Calcium-Ion/new-api", err),

middleware/rate-limit.go

@@ -102,7 +102,10 @@ func GlobalAPIRateLimit() func(c *gin.Context) {
 }
 
 func CriticalRateLimit() func(c *gin.Context) {
-	return rateLimitFactory(common.CriticalRateLimitNum, common.CriticalRateLimitDuration, "CT")
+	if common.CriticalRateLimitEnable {
+		return rateLimitFactory(common.CriticalRateLimitNum, common.CriticalRateLimitDuration, "CT")
+	}
+	return defNext
 }
 
 func DownloadRateLimit() func(c *gin.Context) {

model/channel.go

@@ -688,7 +688,7 @@ func DisableChannelByTag(tag string) error {
 	return err
 }
 
-func EditChannelByTag(tag string, newTag *string, modelMapping *string, models *string, group *string, priority *int64, weight *uint) error {
+func EditChannelByTag(tag string, newTag *string, modelMapping *string, models *string, group *string, priority *int64, weight *uint, paramOverride *string, headerOverride *string) error {
 	updateData := Channel{}
 	shouldReCreateAbilities := false
 	updatedTag := tag
@@ -714,6 +714,12 @@ func EditChannelByTag(tag string, newTag *string, modelMapping *string, models *
 	if weight != nil {
 		updateData.Weight = weight
 	}
+	if paramOverride != nil {
+		updateData.ParamOverride = paramOverride
+	}
+	if headerOverride != nil {
+		updateData.HeaderOverride = headerOverride
+	}
 
 	err := DB.Model(&Channel{}).Where("tag = ?", tag).Updates(updateData).Error
 	if err != nil {

relay/channel/ali/image.go

@@ -98,9 +98,9 @@ func oaiFormEdit2AliImageEdit(c *gin.Context, info *relaycommon.RelayInfo, reque
 		return nil, errors.New("image is required")
 	}
 
-	if len(imageFiles) > 1 {
+	//if len(imageFiles) > 1 {
-		return nil, errors.New("only one image is supported for qwen edit")
+	//	return nil, errors.New("only one image is supported for qwen edit")
-	}
+	//}
 
 	// 获取base64编码的图片
 	var imageBase64s []string

relay/channel/claude/relay-claude.go

@@ -596,6 +596,8 @@ func FormatClaudeResponseInfo(requestMode int, claudeResponse *dto.ClaudeRespons
 			claudeInfo.Usage.PromptTokens = claudeResponse.Message.Usage.InputTokens
 			claudeInfo.Usage.PromptTokensDetails.CachedTokens = claudeResponse.Message.Usage.CacheReadInputTokens
 			claudeInfo.Usage.PromptTokensDetails.CachedCreationTokens = claudeResponse.Message.Usage.CacheCreationInputTokens
+			claudeInfo.Usage.ClaudeCacheCreation5mTokens = claudeResponse.Message.Usage.GetCacheCreation5mTokens()
+			claudeInfo.Usage.ClaudeCacheCreation1hTokens = claudeResponse.Message.Usage.GetCacheCreation1hTokens()
 			claudeInfo.Usage.CompletionTokens = claudeResponse.Message.Usage.OutputTokens
 		} else if claudeResponse.Type == "content_block_delta" {
 			if claudeResponse.Delta.Text != nil {
@@ -740,6 +742,8 @@ func HandleClaudeResponseData(c *gin.Context, info *relaycommon.RelayInfo, claud
 		claudeInfo.Usage.TotalTokens = claudeResponse.Usage.InputTokens + claudeResponse.Usage.OutputTokens
 		claudeInfo.Usage.PromptTokensDetails.CachedTokens = claudeResponse.Usage.CacheReadInputTokens
 		claudeInfo.Usage.PromptTokensDetails.CachedCreationTokens = claudeResponse.Usage.CacheCreationInputTokens
+		claudeInfo.Usage.ClaudeCacheCreation5mTokens = claudeResponse.Usage.GetCacheCreation5mTokens()
+		claudeInfo.Usage.ClaudeCacheCreation1hTokens = claudeResponse.Usage.GetCacheCreation1hTokens()
 	}
 	var responseData []byte
 	switch info.RelayFormat {

relay/channel/openai/relay-openai.go

@@ -122,6 +122,10 @@ func OaiStreamHandler(c *gin.Context, info *relaycommon.RelayInfo, resp *http.Re
 	var usage = &dto.Usage{}
 	var streamItems []string // store stream items
 	var lastStreamData string
+	var secondLastStreamData string // 存储倒数第二个stream data，用于音频模型
+
+	// 检查是否为音频模型
+	isAudioModel := strings.Contains(strings.ToLower(model), "audio")
 
 	helper.StreamScannerHandler(c, resp, info, func(data string) bool {
 		if lastStreamData != "" {
@@ -131,12 +135,35 @@ func OaiStreamHandler(c *gin.Context, info *relaycommon.RelayInfo, resp *http.Re
 			}
 		}
 		if len(data) > 0 {
+			// 对音频模型，保存倒数第二个stream data
+			if isAudioModel && lastStreamData != "" {
+				secondLastStreamData = lastStreamData
+			}
+
 			lastStreamData = data
 			streamItems = append(streamItems, data)
 		}
 		return true
 	})
 
+	// 对音频模型，从倒数第二个stream data中提取usage信息
+	if isAudioModel && secondLastStreamData != "" {
+		var streamResp struct {
+			Usage *dto.Usage `json:"usage"`
+		}
+		err := json.Unmarshal([]byte(secondLastStreamData), &streamResp)
+		if err == nil && streamResp.Usage != nil && service.ValidUsage(streamResp.Usage) {
+			usage = streamResp.Usage
+			containStreamUsage = true
+
+			if common.DebugEnabled {
+				logger.LogDebug(c, fmt.Sprintf("Audio model usage extracted from second last SSE: PromptTokens=%d, CompletionTokens=%d, TotalTokens=%d, InputTokens=%d, OutputTokens=%d",
+					usage.PromptTokens, usage.CompletionTokens, usage.TotalTokens,
+					usage.InputTokens, usage.OutputTokens))
+			}
+		}
+	}
+
 	// 处理最后的响应
 	shouldSendLastResp := true
 	if err := handleLastResponse(lastStreamData, &responseId, &createAt, &systemFingerprint, &model, &usage,

relay/channel/task/ali/adaptor.go

@@ -5,14 +5,17 @@ import (
 	"fmt"
 	"io"
 	"net/http"
+	"strconv"
 	"strings"
 
 	"github.com/QuantumNous/new-api/common"
 	"github.com/QuantumNous/new-api/dto"
+	"github.com/QuantumNous/new-api/logger"
 	"github.com/QuantumNous/new-api/model"
 	"github.com/QuantumNous/new-api/relay/channel"
 	relaycommon "github.com/QuantumNous/new-api/relay/common"
 	"github.com/QuantumNous/new-api/service"
+	"github.com/samber/lo"
 
 	"github.com/gin-gonic/gin"
 	"github.com/pkg/errors"
@@ -108,6 +111,7 @@ type TaskAdaptor struct {
 	ChannelType int
 	apiKey      string
 	baseURL     string
+	aliReq      *AliVideoRequest
 }
 
 func (a *TaskAdaptor) Init(info *relaycommon.RelayInfo) {
@@ -118,6 +122,16 @@ func (a *TaskAdaptor) Init(info *relaycommon.RelayInfo) {
 
 func (a *TaskAdaptor) ValidateRequestAndSetAction(c *gin.Context, info *relaycommon.RelayInfo) (taskErr *dto.TaskError) {
 	// 阿里通义万相支持 JSON 格式，不使用 multipart
+	var taskReq relaycommon.TaskSubmitReq
+	if err := common.UnmarshalBodyReusable(c, &taskReq); err != nil {
+		return service.TaskErrorWrapper(err, "unmarshal_task_request_failed", http.StatusBadRequest)
+	}
+	aliReq, err := a.convertToAliRequest(info, taskReq)
+	if err != nil {
+		return service.TaskErrorWrapper(err, "convert_to_ali_request_failed", http.StatusInternalServerError)
+	}
+	a.aliReq = aliReq
+	logger.LogJson(c, "ali video request body", aliReq)
 	return relaycommon.ValidateMultipartDirect(c, info)
 }
 
@@ -134,13 +148,7 @@ func (a *TaskAdaptor) BuildRequestHeader(c *gin.Context, req *http.Request, info
 }
 
 func (a *TaskAdaptor) BuildRequestBody(c *gin.Context, info *relaycommon.RelayInfo) (io.Reader, error) {
-	var taskReq relaycommon.TaskSubmitReq
+	bodyBytes, err := common.Marshal(a.aliReq)
-	if err := common.UnmarshalBodyReusable(c, &taskReq); err != nil {
-		return nil, errors.Wrap(err, "unmarshal_task_request_failed")
-	}
-	aliReq := a.convertToAliRequest(taskReq)
-
-	bodyBytes, err := common.Marshal(aliReq)
 	if err != nil {
 		return nil, errors.Wrap(err, "marshal_ali_request_failed")
 	}
@@ -148,7 +156,98 @@ func (a *TaskAdaptor) BuildRequestBody(c *gin.Context, info *relaycommon.RelayIn
 	return bytes.NewReader(bodyBytes), nil
 }
 
-func (a *TaskAdaptor) convertToAliRequest(req relaycommon.TaskSubmitReq) *AliVideoRequest {
+var (
+	size480p = []string{
+		"832*480",
+		"480*832",
+		"624*624",
+	}
+	size720p = []string{
+		"1280*720",
+		"720*1280",
+		"960*960",
+		"1088*832",
+		"832*1088",
+	}
+	size1080p = []string{
+		"1920*1080",
+		"1080*1920",
+		"1440*1440",
+		"1632*1248",
+		"1248*1632",
+	}
+)
+
+func sizeToResolution(size string) (string, error) {
+	if lo.Contains(size480p, size) {
+		return "480P", nil
+	} else if lo.Contains(size720p, size) {
+		return "720P", nil
+	} else if lo.Contains(size1080p, size) {
+		return "1080P", nil
+	}
+	return "", fmt.Errorf("invalid size: %s", size)
+}
+
+func ProcessAliOtherRatios(aliReq *AliVideoRequest) (map[string]float64, error) {
+	otherRatios := make(map[string]float64)
+	aliRatios := map[string]map[string]float64{
+		"wan2.5-t2v-preview": {
+			"480P":  1,
+			"720P":  2,
+			"1080P": 1 / 0.3,
+		},
+		"wan2.2-t2v-plus": {
+			"480P":  1,
+			"1080P": 0.7 / 0.14,
+		},
+		"wan2.5-i2v-preview": {
+			"480P":  1,
+			"720P":  2,
+			"1080P": 1 / 0.3,
+		},
+		"wan2.2-i2v-plus": {
+			"480P":  1,
+			"1080P": 0.7 / 0.14,
+		},
+		"wan2.2-kf2v-flash": {
+			"480P":  1,
+			"720P":  2,
+			"1080P": 4.8,
+		},
+		"wan2.2-i2v-flash": {
+			"480P": 1,
+			"720P": 2,
+		},
+		"wan2.2-s2v": {
+			"480P": 1,
+			"720P": 0.9 / 0.5,
+		},
+	}
+	var resolution string
+
+	// size match
+	if aliReq.Parameters.Size != "" {
+		toResolution, err := sizeToResolution(aliReq.Parameters.Size)
+		if err != nil {
+			return nil, err
+		}
+		resolution = toResolution
+	} else {
+		resolution = strings.ToUpper(aliReq.Parameters.Resolution)
+		if !strings.HasSuffix(resolution, "P") {
+			resolution = resolution + "P"
+		}
+	}
+	if otherRatio, ok := aliRatios[aliReq.Model]; ok {
+		if ratio, ok := otherRatio[resolution]; ok {
+			otherRatios[fmt.Sprintf("resolution-%s", resolution)] = ratio
+		}
+	}
+	return otherRatios, nil
+}
+
+func (a *TaskAdaptor) convertToAliRequest(info *relaycommon.RelayInfo, req relaycommon.TaskSubmitReq) (*AliVideoRequest, error) {
 	aliReq := &AliVideoRequest{
 		Model: req.Model,
 		Input: AliVideoInput{
@@ -163,28 +262,53 @@ func (a *TaskAdaptor) convertToAliRequest(req relaycommon.TaskSubmitReq) *AliVid
 
 	// 处理分辨率映射
 	if req.Size != "" {
-		resolution := strings.ToUpper(req.Size)
+		// text to video size must be contained *
-		// 支持 480p, 720p, 1080p 或 480P, 720P, 1080P
+		if strings.Contains(req.Model, "t2v") && !strings.Contains(req.Size, "*") {
-		if !strings.HasSuffix(resolution, "P") {
+			return nil, fmt.Errorf("invalid size: %s, example: %s", req.Size, "1920*1080")
-			resolution = resolution + "P"
+		}
+		if strings.Contains(req.Size, "*") {
+			aliReq.Parameters.Size = req.Size
+		} else {
+			resolution := strings.ToUpper(req.Size)
+			// 支持 480p, 720p, 1080p 或 480P, 720P, 1080P
+			if !strings.HasSuffix(resolution, "P") {
+				resolution = resolution + "P"
+			}
+			aliReq.Parameters.Resolution = resolution
 		}
-		aliReq.Parameters.Resolution = resolution
 	} else {
 		// 根据模型设置默认分辨率
-		if strings.HasPrefix(req.Model, "wan2.5") {
+		if strings.Contains(req.Model, "t2v") { // image to video
-			aliReq.Parameters.Resolution = "1080P"
+			if strings.HasPrefix(req.Model, "wan2.5") {
-		} else if strings.HasPrefix(req.Model, "wan2.2-i2v-flash") {
+				aliReq.Parameters.Size = "1920*1080"
-			aliReq.Parameters.Resolution = "720P"
+			} else if strings.HasPrefix(req.Model, "wan2.2") {
-		} else if strings.HasPrefix(req.Model, "wan2.2-i2v-plus") {
+				aliReq.Parameters.Size = "1920*1080"
-			aliReq.Parameters.Resolution = "1080P"
+			} else {
+				aliReq.Parameters.Size = "1280*720"
+			}
 		} else {
-			aliReq.Parameters.Resolution = "720P"
+			if strings.HasPrefix(req.Model, "wan2.5") {
+				aliReq.Parameters.Resolution = "1080P"
+			} else if strings.HasPrefix(req.Model, "wan2.2-i2v-flash") {
+				aliReq.Parameters.Resolution = "720P"
+			} else if strings.HasPrefix(req.Model, "wan2.2-i2v-plus") {
+				aliReq.Parameters.Resolution = "1080P"
+			} else {
+				aliReq.Parameters.Resolution = "720P"
+			}
 		}
 	}
 
 	// 处理时长
 	if req.Duration > 0 {
 		aliReq.Parameters.Duration = req.Duration
+	} else if req.Seconds != "" {
+		seconds, err := strconv.Atoi(req.Seconds)
+		if err != nil {
+			return nil, errors.Wrap(err, "convert seconds to int failed")
+		} else {
+			aliReq.Parameters.Duration = seconds
+		}
 	} else {
 		aliReq.Parameters.Duration = 5 // 默认5秒
 	}
@@ -192,11 +316,32 @@ func (a *TaskAdaptor) convertToAliRequest(req relaycommon.TaskSubmitReq) *AliVid
 	// 从 metadata 中提取额外参数
 	if req.Metadata != nil {
 		if metadataBytes, err := common.Marshal(req.Metadata); err == nil {
-			_ = common.Unmarshal(metadataBytes, aliReq)
+			err = common.Unmarshal(metadataBytes, aliReq)
+			if err != nil {
+				return nil, errors.Wrap(err, "unmarshal metadata failed")
+			}
+		} else {
+			return nil, errors.Wrap(err, "marshal metadata failed")
 		}
 	}
 
-	return aliReq
+	if aliReq.Model != req.Model {
+		return nil, errors.New("can't change model with metadata")
+	}
+
+	info.PriceData.OtherRatios = map[string]float64{
+		"seconds": float64(aliReq.Parameters.Duration),
+	}
+
+	ratios, err := ProcessAliOtherRatios(aliReq)
+	if err != nil {
+		return nil, err
+	}
+	for s, f := range ratios {
+		info.PriceData.OtherRatios[s] = f
+	}
+
+	return aliReq, nil
 }
 
 // DoRequest delegates to common helper

relay/channel/task/jimeng/adaptor.go

@@ -406,12 +406,15 @@ func (a *TaskAdaptor) convertToRequestPayload(req *relaycommon.TaskSubmitReq) (*
 	// 即梦视频3.0 ReqKey转换
 	// https://www.volcengine.com/docs/85621/1792707
 	if strings.Contains(r.ReqKey, "jimeng_v30") {
-		if len(req.Images) > 1 {
+		if r.ReqKey == "jimeng_v30_pro" {
+			// 3.0 pro只有固定的jimeng_ti2v_v30_pro
+			r.ReqKey = "jimeng_ti2v_v30_pro"
+		} else if len(req.Images) > 1 {
 			// 多张图片：首尾帧生成
-			r.ReqKey = strings.Replace(r.ReqKey, "jimeng_v30", "jimeng_i2v_first_tail_v30", 1)
+			r.ReqKey = strings.TrimSuffix(strings.Replace(r.ReqKey, "jimeng_v30", "jimeng_i2v_first_tail_v30", 1), "p")
 		} else if len(req.Images) == 1 {
 			// 单张图片：图生视频
-			r.ReqKey = strings.Replace(r.ReqKey, "jimeng_v30", "jimeng_i2v_first_v30", 1)
+			r.ReqKey = strings.TrimSuffix(strings.Replace(r.ReqKey, "jimeng_v30", "jimeng_i2v_first_v30", 1), "p")
 		} else {
 			// 无图片：文生视频
 			r.ReqKey = strings.Replace(r.ReqKey, "jimeng_v30", "jimeng_t2v_v30", 1)

relay/common/relay_utils.go

@@ -121,6 +121,7 @@ func ValidateMultipartDirect(c *gin.Context, info *RelayInfo) *dto.TaskError {
 
 	prompt = req.Prompt
 	model = req.Model
+	size = req.Size
 	seconds, _ = strconv.Atoi(req.Seconds)
 	if seconds == 0 {
 		seconds = req.Duration

relay/helper/price.go

@@ -13,6 +13,9 @@ import (
 	"github.com/gin-gonic/gin"
 )
 
+// https://docs.claude.com/en/docs/build-with-claude/prompt-caching#1-hour-cache-duration
+const claudeCacheCreation1hMultiplier = 6 / 3.75
+
 // HandleGroupRatio checks for "auto_group" in the context and updates the group ratio and relayInfo.UsingGroup if present
 func HandleGroupRatio(ctx *gin.Context, relayInfo *relaycommon.RelayInfo) types.GroupRatioInfo {
 	groupRatioInfo := types.GroupRatioInfo{
@@ -53,6 +56,8 @@ func ModelPriceHelper(c *gin.Context, info *relaycommon.RelayInfo, promptTokens
 	var cacheRatio float64
 	var imageRatio float64
 	var cacheCreationRatio float64
+	var cacheCreationRatio5m float64
+	var cacheCreationRatio1h float64
 	var audioRatio float64
 	var audioCompletionRatio float64
 	var freeModel bool
@@ -76,6 +81,9 @@ func ModelPriceHelper(c *gin.Context, info *relaycommon.RelayInfo, promptTokens
 		completionRatio = ratio_setting.GetCompletionRatio(info.OriginModelName)
 		cacheRatio, _ = ratio_setting.GetCacheRatio(info.OriginModelName)
 		cacheCreationRatio, _ = ratio_setting.GetCreateCacheRatio(info.OriginModelName)
+		cacheCreationRatio5m = cacheCreationRatio
+		// 固定1h和5min缓存写入价格的比例
+		cacheCreationRatio1h = cacheCreationRatio * claudeCacheCreation1hMultiplier
 		imageRatio, _ = ratio_setting.GetImageRatio(info.OriginModelName)
 		audioRatio = ratio_setting.GetAudioRatio(info.OriginModelName)
 		audioCompletionRatio = ratio_setting.GetAudioCompletionRatio(info.OriginModelName)
@@ -116,6 +124,8 @@ func ModelPriceHelper(c *gin.Context, info *relaycommon.RelayInfo, promptTokens
 		AudioRatio:           audioRatio,
 		AudioCompletionRatio: audioCompletionRatio,
 		CacheCreationRatio:   cacheCreationRatio,
+		CacheCreation5mRatio: cacheCreationRatio5m,
+		CacheCreation1hRatio: cacheCreationRatio1h,
 		QuotaToPreConsume:    preConsumedQuota,
 	}
 

service/log_info_generate.go

@@ -92,11 +92,23 @@ func GenerateAudioOtherInfo(ctx *gin.Context, relayInfo *relaycommon.RelayInfo,
 }
 
 func GenerateClaudeOtherInfo(ctx *gin.Context, relayInfo *relaycommon.RelayInfo, modelRatio, groupRatio, completionRatio float64,
-	cacheTokens int, cacheRatio float64, cacheCreationTokens int, cacheCreationRatio float64, modelPrice float64, userGroupRatio float64) map[string]interface{} {
+	cacheTokens int, cacheRatio float64,
+	cacheCreationTokens int, cacheCreationRatio float64,
+	cacheCreationTokens5m int, cacheCreationRatio5m float64,
+	cacheCreationTokens1h int, cacheCreationRatio1h float64,
+	modelPrice float64, userGroupRatio float64) map[string]interface{} {
 	info := GenerateTextOtherInfo(ctx, relayInfo, modelRatio, groupRatio, completionRatio, cacheTokens, cacheRatio, modelPrice, userGroupRatio)
 	info["claude"] = true
 	info["cache_creation_tokens"] = cacheCreationTokens
 	info["cache_creation_ratio"] = cacheCreationRatio
+	if cacheCreationTokens5m != 0 {
+		info["cache_creation_tokens_5m"] = cacheCreationTokens5m
+		info["cache_creation_ratio_5m"] = cacheCreationRatio5m
+	}
+	if cacheCreationTokens1h != 0 {
+		info["cache_creation_tokens_1h"] = cacheCreationTokens1h
+		info["cache_creation_ratio_1h"] = cacheCreationRatio1h
+	}
 	return info
 }
 

service/quota.go

@@ -251,7 +251,11 @@ func PostClaudeConsumeQuota(ctx *gin.Context, relayInfo *relaycommon.RelayInfo,
 	cacheTokens := usage.PromptTokensDetails.CachedTokens
 
 	cacheCreationRatio := relayInfo.PriceData.CacheCreationRatio
+	cacheCreationRatio5m := relayInfo.PriceData.CacheCreation5mRatio
+	cacheCreationRatio1h := relayInfo.PriceData.CacheCreation1hRatio
 	cacheCreationTokens := usage.PromptTokensDetails.CachedCreationTokens
+	cacheCreationTokens5m := usage.ClaudeCacheCreation5mTokens
+	cacheCreationTokens1h := usage.ClaudeCacheCreation1hTokens
 
 	if relayInfo.ChannelType == constant.ChannelTypeOpenRouter {
 		promptTokens -= cacheTokens
@@ -269,7 +273,12 @@ func PostClaudeConsumeQuota(ctx *gin.Context, relayInfo *relaycommon.RelayInfo,
 	if !relayInfo.PriceData.UsePrice {
 		calculateQuota = float64(promptTokens)
 		calculateQuota += float64(cacheTokens) * cacheRatio
-		calculateQuota += float64(cacheCreationTokens) * cacheCreationRatio
+		calculateQuota += float64(cacheCreationTokens5m) * cacheCreationRatio5m
+		calculateQuota += float64(cacheCreationTokens1h) * cacheCreationRatio1h
+		remainingCacheCreationTokens := cacheCreationTokens - cacheCreationTokens5m - cacheCreationTokens1h
+		if remainingCacheCreationTokens > 0 {
+			calculateQuota += float64(remainingCacheCreationTokens) * cacheCreationRatio
+		}
 		calculateQuota += float64(completionTokens) * completionRatio
 		calculateQuota = calculateQuota * groupRatio * modelRatio
 	} else {
@@ -322,7 +331,11 @@ func PostClaudeConsumeQuota(ctx *gin.Context, relayInfo *relaycommon.RelayInfo,
 	}
 
 	other := GenerateClaudeOtherInfo(ctx, relayInfo, modelRatio, groupRatio, completionRatio,
-		cacheTokens, cacheRatio, cacheCreationTokens, cacheCreationRatio, modelPrice, relayInfo.PriceData.GroupRatioInfo.GroupSpecialRatio)
+		cacheTokens, cacheRatio,
+		cacheCreationTokens, cacheCreationRatio,
+		cacheCreationTokens5m, cacheCreationRatio5m,
+		cacheCreationTokens1h, cacheCreationRatio1h,
+		modelPrice, relayInfo.PriceData.GroupRatioInfo.GroupSpecialRatio)
 	model.RecordConsumeLog(ctx, relayInfo.UserId, model.RecordConsumeLogParams{
 		ChannelId:        relayInfo.ChannelId,
 		PromptTokens:     promptTokens,

types/price_data.go

@@ -15,6 +15,8 @@ type PriceData struct {
 	CompletionRatio      float64
 	CacheRatio           float64
 	CacheCreationRatio   float64
+	CacheCreation5mRatio float64
+	CacheCreation1hRatio float64
 	ImageRatio           float64
 	AudioRatio           float64
 	AudioCompletionRatio float64
@@ -31,5 +33,5 @@ type PerCallPriceData struct {
 }
 
 func (p PriceData) ToSetting() string {
-	return fmt.Sprintf("ModelPrice: %f, ModelRatio: %f, CompletionRatio: %f, CacheRatio: %f, GroupRatio: %f, UsePrice: %t, CacheCreationRatio: %f, QuotaToPreConsume: %d, ImageRatio: %f, AudioRatio: %f, AudioCompletionRatio: %f", p.ModelPrice, p.ModelRatio, p.CompletionRatio, p.CacheRatio, p.GroupRatioInfo.GroupRatio, p.UsePrice, p.CacheCreationRatio, p.QuotaToPreConsume, p.ImageRatio, p.AudioRatio, p.AudioCompletionRatio)
+	return fmt.Sprintf("ModelPrice: %f, ModelRatio: %f, CompletionRatio: %f, CacheRatio: %f, GroupRatio: %f, UsePrice: %t, CacheCreationRatio: %f, CacheCreation5mRatio: %f, CacheCreation1hRatio: %f, QuotaToPreConsume: %d, ImageRatio: %f, AudioRatio: %f, AudioCompletionRatio: %f", p.ModelPrice, p.ModelRatio, p.CompletionRatio, p.CacheRatio, p.GroupRatioInfo.GroupRatio, p.UsePrice, p.CacheCreationRatio, p.CacheCreation5mRatio, p.CacheCreation1hRatio, p.QuotaToPreConsume, p.ImageRatio, p.AudioRatio, p.AudioCompletionRatio)
 }

web/src/components/table/channels/modals/EditTagModal.jsx

@@ -45,6 +45,7 @@ import {
   IconBookmark,
   IconUser,
   IconCode,
+  IconSetting,
 } from '@douyinfe/semi-icons';
 import { getChannelModels } from '../../../../helpers';
 import { useTranslation } from 'react-i18next';
@@ -69,6 +70,8 @@ const EditTagModal = (props) => {
     model_mapping: null,
     groups: [],
     models: [],
+    param_override: null,
+    header_override: null,
   };
   const [inputs, setInputs] = useState(originInputs);
   const formApiRef = useRef(null);
@@ -190,12 +193,48 @@ const EditTagModal = (props) => {
     if (formVals.models && formVals.models.length > 0) {
       data.models = formVals.models.join(',');
     }
+    if (
+      formVals.param_override !== undefined &&
+      formVals.param_override !== null
+    ) {
+      if (typeof formVals.param_override !== 'string') {
+        showInfo('参数覆盖必须是合法的 JSON 格式！');
+        setLoading(false);
+        return;
+      }
+      const trimmedParamOverride = formVals.param_override.trim();
+      if (trimmedParamOverride !== '' && !verifyJSON(trimmedParamOverride)) {
+        showInfo('参数覆盖必须是合法的 JSON 格式！');
+        setLoading(false);
+        return;
+      }
+      data.param_override = trimmedParamOverride;
+    }
+    if (
+      formVals.header_override !== undefined &&
+      formVals.header_override !== null
+    ) {
+      if (typeof formVals.header_override !== 'string') {
+        showInfo('请求头覆盖必须是合法的 JSON 格式！');
+        setLoading(false);
+        return;
+      }
+      const trimmedHeaderOverride = formVals.header_override.trim();
+      if (trimmedHeaderOverride !== '' && !verifyJSON(trimmedHeaderOverride)) {
+        showInfo('请求头覆盖必须是合法的 JSON 格式！');
+        setLoading(false);
+        return;
+      }
+      data.header_override = trimmedHeaderOverride;
+    }
     data.new_tag = formVals.new_tag;
     if (
       data.model_mapping === undefined &&
       data.groups === undefined &&
       data.models === undefined &&
-      data.new_tag === undefined
+      data.new_tag === undefined &&
+      data.param_override === undefined &&
+      data.header_override === undefined
     ) {
       showWarning('没有任何修改！');
       setLoading(false);
@@ -491,6 +530,157 @@ const EditTagModal = (props) => {
                 </div>
               </Card>
 
+              <Card className='!rounded-2xl shadow-sm border-0 mb-6'>
+                {/* Header: Advanced Settings */}
+                <div className='flex items-center mb-2'>
+                  <Avatar size='small' color='orange' className='mr-2 shadow-md'>
+                    <IconSetting size={16} />
+                  </Avatar>
+                  <div>
+                    <Text className='text-lg font-medium'>{t('高级设置')}</Text>
+                    <div className='text-xs text-gray-600'>
+                      {t('渠道的高级配置选项')}
+                    </div>
+                  </div>
+                </div>
+
+                <div className='space-y-4'>
+                  <Form.TextArea
+                    field='param_override'
+                    label={t('参数覆盖')}
+                    placeholder={
+                      t(
+                        '此项可选，用于覆盖请求参数。不支持覆盖 stream 参数',
+                      ) +
+                      '\n' +
+                      t('旧格式（直接覆盖）：') +
+                      '\n{\n  "temperature": 0,\n  "max_tokens": 1000\n}' +
+                      '\n\n' +
+                      t('新格式（支持条件判断与json自定义）：') +
+                      '\n{\n  "operations": [\n    {\n      "path": "temperature",\n      "mode": "set",\n      "value": 0.7,\n      "conditions": [\n        {\n          "path": "model",\n          "mode": "prefix",\n          "value": "gpt"\n        }\n      ]\n    }\n  ]\n}'
+                    }
+                    autosize
+                    showClear
+                    onChange={(value) =>
+                      handleInputChange('param_override', value)
+                    }
+                    extraText={
+                      <div className='flex gap-2 flex-wrap'>
+                        <Text
+                          className='!text-semi-color-primary cursor-pointer'
+                          onClick={() =>
+                            handleInputChange(
+                              'param_override',
+                              JSON.stringify({ temperature: 0 }, null, 2),
+                            )
+                          }
+                        >
+                          {t('旧格式模板')}
+                        </Text>
+                        <Text
+                          className='!text-semi-color-primary cursor-pointer'
+                          onClick={() =>
+                            handleInputChange(
+                              'param_override',
+                              JSON.stringify(
+                                {
+                                  operations: [
+                                    {
+                                      path: 'temperature',
+                                      mode: 'set',
+                                      value: 0.7,
+                                      conditions: [
+                                        {
+                                          path: 'model',
+                                          mode: 'prefix',
+                                          value: 'gpt',
+                                        },
+                                      ],
+                                      logic: 'AND',
+                                    },
+                                  ],
+                                },
+                                null,
+                                2,
+                              ),
+                            )
+                          }
+                        >
+                          {t('新格式模板')}
+                        </Text>
+                        <Text
+                          className='!text-semi-color-primary cursor-pointer'
+                          onClick={() =>
+                            handleInputChange('param_override', null)
+                          }
+                        >
+                          {t('不更改')}
+                        </Text>
+                      </div>
+                    }
+                  />
+
+                  <Form.TextArea
+                    field='header_override'
+                    label={t('请求头覆盖')}
+                    placeholder={
+                      t('此项可选，用于覆盖请求头参数') +
+                      '\n' +
+                      t('格式示例：') +
+                      '\n{\n  "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/139.0.0.0 Safari/537.36 Edg/139.0.0.0",\n  "Authorization": "Bearer {api_key}"\n}'
+                    }
+                    autosize
+                    showClear
+                    onChange={(value) =>
+                      handleInputChange('header_override', value)
+                    }
+                    extraText={
+                      <div className='flex flex-col gap-1'>
+                        <div className='flex gap-2 flex-wrap items-center'>
+                          <Text
+                            className='!text-semi-color-primary cursor-pointer'
+                            onClick={() =>
+                              handleInputChange(
+                                'header_override',
+                                JSON.stringify(
+                                  {
+                                    'User-Agent':
+                                      'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/139.0.0.0 Safari/537.36 Edg/139.0.0.0',
+                                    Authorization: 'Bearer {api_key}',
+                                  },
+                                  null,
+                                  2,
+                                ),
+                              )
+                            }
+                          >
+                            {t('填入模板')}
+                          </Text>
+                          <Text
+                            className='!text-semi-color-primary cursor-pointer'
+                            onClick={() =>
+                              handleInputChange('header_override', null)
+                            }
+                          >
+                            {t('不更改')}
+                          </Text>
+                        </div>
+                        <div>
+                          <Text type='tertiary' size='small'>
+                            {t('支持变量：')}
+                          </Text>
+                          <div className='text-xs text-tertiary ml-2'>
+                            <div>
+                              {t('渠道密钥')}: {'{api_key}'}
+                            </div>
+                          </div>
+                        </div>
+                      </div>
+                    }
+                  />
+                </div>
+              </Card>
+
               <Card className='!rounded-2xl shadow-sm border-0'>
                 {/* Header: Group Settings */}
                 <div className='flex items-center mb-2'>

web/src/components/table/tokens/modals/EditTokenModal.jsx

@@ -66,9 +66,9 @@ const EditTokenModal = (props) => {
 
   const getInitValues = () => ({
     name: '',
-    remain_quota: 500000,
+    remain_quota: 0,
     expired_time: -1,
-    unlimited_quota: false,
+    unlimited_quota: true,
     model_limits_enabled: false,
     model_limits: [],
     allow_ips: '',

web/src/components/table/usage-logs/UsageLogsColumnDefs.jsx

@@ -551,6 +551,10 @@ export const getLogsColumns = ({
               other.cache_ratio || 1.0,
               other.cache_creation_tokens || 0,
               other.cache_creation_ratio || 1.0,
+              other.cache_creation_tokens_5m || 0,
+              other.cache_creation_ratio_5m || other.cache_creation_ratio || 1.0,
+              other.cache_creation_tokens_1h || 0,
+              other.cache_creation_ratio_1h || other.cache_creation_ratio || 1.0,
               false,
               1.0,
               other?.is_system_prompt_overwritten,
@@ -565,6 +569,10 @@ export const getLogsColumns = ({
               other.cache_ratio || 1.0,
               0,
               1.0,
+              0,
+              1.0,
+              0,
+              1.0,
               false,
               1.0,
               other?.is_system_prompt_overwritten,

web/src/helpers/base64.js

@@ -0,0 +1,56 @@
+/*
+Copyright (C) 2025 QuantumNous
+
+This program is free software: you can redistribute it and/or modify
+it under the terms of the GNU Affero General Public License as
+published by the Free Software Foundation, either version 3 of the
+License, or (at your option) any later version.
+
+This program is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+GNU Affero General Public License for more details.
+
+You should have received a copy of the GNU Affero General Public License
+along with this program. If not, see <https://www.gnu.org/licenses/>.
+
+For commercial licensing, please contact support@quantumnous.com
+*/
+
+const toBinaryString = (text) => {
+  if (typeof TextEncoder !== 'undefined') {
+    const bytes = new TextEncoder().encode(text);
+    let binary = '';
+
+    bytes.forEach((byte) => {
+      binary += String.fromCharCode(byte);
+    });
+
+    return binary;
+  }
+
+  return encodeURIComponent(text).replace(/%([0-9A-F]{2})/g, (_, hex) =>
+    String.fromCharCode(parseInt(hex, 16)),
+  );
+};
+
+export const encodeToBase64 = (value) => {
+  const input = value == null ? '' : String(value);
+
+  if (typeof window === 'undefined') {
+    if (typeof Buffer !== 'undefined') {
+      return Buffer.from(input, 'utf-8').toString('base64');
+    }
+    if (
+      typeof globalThis !== 'undefined' &&
+      typeof globalThis.btoa === 'function'
+    ) {
+      return globalThis.btoa(toBinaryString(input));
+    }
+    throw new Error(
+      'Base64 encoding is unavailable in the current environment',
+    );
+  }
+
+  return window.btoa(toBinaryString(input));
+};

web/src/helpers/index.js

@@ -20,6 +20,7 @@ For commercial licensing, please contact support@quantumnous.com
 export * from './history';
 export * from './auth';
 export * from './utils';
+export * from './base64';
 export * from './api';
 export * from './render';
 export * from './log';

web/src/helpers/render.jsx

@@ -1046,6 +1046,10 @@ function renderPriceSimpleCore({
   cacheRatio = 1.0,
   cacheCreationTokens = 0,
   cacheCreationRatio = 1.0,
+  cacheCreationTokens5m = 0,
+  cacheCreationRatio5m = 1.0,
+  cacheCreationTokens1h = 0,
+  cacheCreationRatio1h = 1.0,
   image = false,
   imageRatio = 1.0,
   isSystemPromptOverride = false,
@@ -1064,17 +1068,40 @@ function renderPriceSimpleCore({
     });
   }
 
+  const hasSplitCacheCreation =
+    cacheCreationTokens5m > 0 || cacheCreationTokens1h > 0;
+
+  const shouldShowLegacyCacheCreation =
+    !hasSplitCacheCreation && cacheCreationTokens !== 0;
+
+  const shouldShowCache = cacheTokens !== 0;
+  const shouldShowCacheCreation5m =
+    hasSplitCacheCreation && cacheCreationTokens5m > 0;
+  const shouldShowCacheCreation1h =
+    hasSplitCacheCreation && cacheCreationTokens1h > 0;
+
   const parts = [];
   // base: model ratio
   parts.push(i18next.t('模型: {{ratio}}'));
 
   // cache part (label differs when with image)
-  if (cacheTokens !== 0) {
+  if (shouldShowCache) {
     parts.push(i18next.t('缓存: {{cacheRatio}}'));
   }
 
-  // cache creation part (Claude specific if passed)
+  if (hasSplitCacheCreation) {
-  if (cacheCreationTokens !== 0) {
+    if (shouldShowCacheCreation5m && shouldShowCacheCreation1h) {
+      parts.push(
+        i18next.t(
+          '缓存创建: 5m {{cacheCreationRatio5m}} / 1h {{cacheCreationRatio1h}}',
+        ),
+      );
+    } else if (shouldShowCacheCreation5m) {
+      parts.push(i18next.t('缓存创建: 5m {{cacheCreationRatio5m}}'));
+    } else if (shouldShowCacheCreation1h) {
+      parts.push(i18next.t('缓存创建: 1h {{cacheCreationRatio1h}}'));
+    }
+  } else if (shouldShowLegacyCacheCreation) {
     parts.push(i18next.t('缓存创建: {{cacheCreationRatio}}'));
   }
 
@@ -1091,6 +1118,8 @@ function renderPriceSimpleCore({
     groupRatio: finalGroupRatio,
     cacheRatio: cacheRatio,
     cacheCreationRatio: cacheCreationRatio,
+    cacheCreationRatio5m: cacheCreationRatio5m,
+    cacheCreationRatio1h: cacheCreationRatio1h,
     imageRatio: imageRatio,
   });
 
@@ -1450,6 +1479,10 @@ export function renderModelPriceSimple(
   cacheRatio = 1.0,
   cacheCreationTokens = 0,
   cacheCreationRatio = 1.0,
+  cacheCreationTokens5m = 0,
+  cacheCreationRatio5m = 1.0,
+  cacheCreationTokens1h = 0,
+  cacheCreationRatio1h = 1.0,
   image = false,
   imageRatio = 1.0,
   isSystemPromptOverride = false,
@@ -1464,6 +1497,10 @@ export function renderModelPriceSimple(
     cacheRatio,
     cacheCreationTokens,
     cacheCreationRatio,
+    cacheCreationTokens5m,
+    cacheCreationRatio5m,
+    cacheCreationTokens1h,
+    cacheCreationRatio1h,
     image,
     imageRatio,
     isSystemPromptOverride,
@@ -1681,6 +1718,10 @@ export function renderClaudeModelPrice(
   cacheRatio = 1.0,
   cacheCreationTokens = 0,
   cacheCreationRatio = 1.0,
+  cacheCreationTokens5m = 0,
+  cacheCreationRatio5m = 1.0,
+  cacheCreationTokens1h = 0,
+  cacheCreationRatio1h = 1.0,
 ) {
   const { ratio: effectiveGroupRatio, label: ratioLabel } = getEffectiveRatio(
     groupRatio,
@@ -1710,20 +1751,121 @@ export function renderClaudeModelPrice(
     const completionRatioValue = completionRatio || 0;
     const inputRatioPrice = modelRatio * 2.0;
     const completionRatioPrice = modelRatio * 2.0 * completionRatioValue;
-    let cacheRatioPrice = (modelRatio * 2.0 * cacheRatio).toFixed(2);
+    const cacheRatioPrice = modelRatio * 2.0 * cacheRatio;
-    let cacheCreationRatioPrice = modelRatio * 2.0 * cacheCreationRatio;
+    const cacheCreationRatioPrice = modelRatio * 2.0 * cacheCreationRatio;
+    const cacheCreationRatioPrice5m = modelRatio * 2.0 * cacheCreationRatio5m;
+    const cacheCreationRatioPrice1h = modelRatio * 2.0 * cacheCreationRatio1h;
+
+    const hasSplitCacheCreation =
+      cacheCreationTokens5m > 0 || cacheCreationTokens1h > 0;
+
+    const shouldShowCache = cacheTokens > 0;
+    const shouldShowLegacyCacheCreation =
+      !hasSplitCacheCreation && cacheCreationTokens > 0;
+    const shouldShowCacheCreation5m =
+      hasSplitCacheCreation && cacheCreationTokens5m > 0;
+    const shouldShowCacheCreation1h =
+      hasSplitCacheCreation && cacheCreationTokens1h > 0;
 
     // Calculate effective input tokens (non-cached + cached with ratio applied + cache creation with ratio applied)
     const nonCachedTokens = inputTokens;
     const effectiveInputTokens =
       nonCachedTokens +
       cacheTokens * cacheRatio +
-      cacheCreationTokens * cacheCreationRatio;
+      cacheCreationTokens * cacheCreationRatio +
+      cacheCreationTokens5m * cacheCreationRatio5m +
+      cacheCreationTokens1h * cacheCreationRatio1h;
 
     let price =
       (effectiveInputTokens / 1000000) * inputRatioPrice * groupRatio +
       (completionTokens / 1000000) * completionRatioPrice * groupRatio;
 
+    const inputUnitPrice = inputRatioPrice * rate;
+    const completionUnitPrice = completionRatioPrice * rate;
+    const cacheUnitPrice = cacheRatioPrice * rate;
+    const cacheCreationUnitPrice = cacheCreationRatioPrice * rate;
+    const cacheCreationUnitPrice5m = cacheCreationRatioPrice5m * rate;
+    const cacheCreationUnitPrice1h = cacheCreationRatioPrice1h * rate;
+    const cacheCreationUnitPriceTotal =
+      cacheCreationUnitPrice5m + cacheCreationUnitPrice1h;
+
+    const breakdownSegments = [
+      i18next.t('提示 {{input}} tokens / 1M tokens * {{symbol}}{{price}}', {
+        input: inputTokens,
+        symbol,
+        price: inputUnitPrice.toFixed(6),
+      }),
+    ];
+
+    if (shouldShowCache) {
+      breakdownSegments.push(
+        i18next.t(
+          '缓存 {{tokens}} tokens / 1M tokens * {{symbol}}{{price}} (倍率: {{ratio}})',
+          {
+            tokens: cacheTokens,
+            symbol,
+            price: cacheUnitPrice.toFixed(6),
+            ratio: cacheRatio,
+          },
+        ),
+      );
+    }
+
+    if (shouldShowLegacyCacheCreation) {
+      breakdownSegments.push(
+        i18next.t(
+          '缓存创建 {{tokens}} tokens / 1M tokens * {{symbol}}{{price}} (倍率: {{ratio}})',
+          {
+            tokens: cacheCreationTokens,
+            symbol,
+            price: cacheCreationUnitPrice.toFixed(6),
+            ratio: cacheCreationRatio,
+          },
+        ),
+      );
+    }
+
+    if (shouldShowCacheCreation5m) {
+      breakdownSegments.push(
+        i18next.t(
+          '5m缓存创建 {{tokens}} tokens / 1M tokens * {{symbol}}{{price}} (倍率: {{ratio}})',
+          {
+            tokens: cacheCreationTokens5m,
+            symbol,
+            price: cacheCreationUnitPrice5m.toFixed(6),
+            ratio: cacheCreationRatio5m,
+          },
+        ),
+      );
+    }
+
+    if (shouldShowCacheCreation1h) {
+      breakdownSegments.push(
+        i18next.t(
+          '1h缓存创建 {{tokens}} tokens / 1M tokens * {{symbol}}{{price}} (倍率: {{ratio}})',
+          {
+            tokens: cacheCreationTokens1h,
+            symbol,
+            price: cacheCreationUnitPrice1h.toFixed(6),
+            ratio: cacheCreationRatio1h,
+          },
+        ),
+      );
+    }
+
+    breakdownSegments.push(
+      i18next.t(
+        '补全 {{completion}} tokens / 1M tokens * {{symbol}}{{price}}',
+        {
+          completion: completionTokens,
+          symbol,
+          price: completionUnitPrice.toFixed(6),
+        },
+      ),
+    );
+
+    const breakdownText = breakdownSegments.join(' + ');
+
     return (
       <>
         <article>
@@ -1744,7 +1886,7 @@ export function renderClaudeModelPrice(
               },
             )}
           </p>
-          {cacheTokens > 0 && (
+          {shouldShowCache && (
             <p>
               {i18next.t(
                 '缓存价格：{{symbol}}{{price}} * {{ratio}} = {{symbol}}{{total}} / 1M tokens (缓存倍率: {{cacheRatio}})',
@@ -1752,13 +1894,13 @@ export function renderClaudeModelPrice(
                   symbol: symbol,
                   price: (inputRatioPrice * rate).toFixed(6),
                   ratio: cacheRatio,
-                  total: (cacheRatioPrice * rate).toFixed(2),
+                  total: cacheUnitPrice.toFixed(6),
                   cacheRatio: cacheRatio,
                 },
               )}
             </p>
           )}
-          {cacheCreationTokens > 0 && (
+          {shouldShowLegacyCacheCreation && (
             <p>
               {i18next.t(
                 '缓存创建价格：{{symbol}}{{price}} * {{ratio}} = {{symbol}}{{total}} / 1M tokens (缓存创建倍率: {{cacheCreationRatio}})',
@@ -1766,49 +1908,65 @@ export function renderClaudeModelPrice(
                   symbol: symbol,
                   price: (inputRatioPrice * rate).toFixed(6),
                   ratio: cacheCreationRatio,
-                  total: (cacheCreationRatioPrice * rate).toFixed(6),
+                  total: cacheCreationUnitPrice.toFixed(6),
                   cacheCreationRatio: cacheCreationRatio,
                 },
               )}
             </p>
           )}
+          {shouldShowCacheCreation5m && (
+            <p>
+              {i18next.t(
+                '5m缓存创建价格：{{symbol}}{{price}} * {{ratio}} = {{symbol}}{{total}} / 1M tokens (5m缓存创建倍率: {{cacheCreationRatio5m}})',
+                {
+                  symbol: symbol,
+                  price: (inputRatioPrice * rate).toFixed(6),
+                  ratio: cacheCreationRatio5m,
+                  total: cacheCreationUnitPrice5m.toFixed(6),
+                  cacheCreationRatio5m: cacheCreationRatio5m,
+                },
+              )}
+            </p>
+          )}
+          {shouldShowCacheCreation1h && (
+            <p>
+              {i18next.t(
+                '1h缓存创建价格：{{symbol}}{{price}} * {{ratio}} = {{symbol}}{{total}} / 1M tokens (1h缓存创建倍率: {{cacheCreationRatio1h}})',
+                {
+                  symbol: symbol,
+                  price: (inputRatioPrice * rate).toFixed(6),
+                  ratio: cacheCreationRatio1h,
+                  total: cacheCreationUnitPrice1h.toFixed(6),
+                  cacheCreationRatio1h: cacheCreationRatio1h,
+                },
+              )}
+            </p>
+          )}
+          {shouldShowCacheCreation5m && shouldShowCacheCreation1h && (
+            <p>
+              {i18next.t(
+                '缓存创建价格合计：5m {{symbol}}{{five}} + 1h {{symbol}}{{one}} = {{symbol}}{{total}} / 1M tokens',
+                {
+                  symbol: symbol,
+                  five: cacheCreationUnitPrice5m.toFixed(6),
+                  one: cacheCreationUnitPrice1h.toFixed(6),
+                  total: cacheCreationUnitPriceTotal.toFixed(6),
+                },
+              )}
+            </p>
+          )}
           <p></p>
           <p>
-            {cacheTokens > 0 || cacheCreationTokens > 0
+            {i18next.t(
-              ? i18next.t(
+              '{{breakdown}} * {{ratioType}} {{ratio}} = {{symbol}}{{total}}',
-                  '提示 {{nonCacheInput}} tokens / 1M tokens * {{symbol}}{{price}} + 缓存 {{cacheInput}} tokens / 1M tokens * {{symbol}}{{cachePrice}} + 缓存创建 {{cacheCreationInput}} tokens / 1M tokens * {{symbol}}{{cacheCreationPrice}} + 补全 {{completion}} tokens / 1M tokens * {{symbol}}{{compPrice}} * {{ratioType}} {{ratio}} = {{symbol}}{{total}}',
+              {
-                  {
+                breakdown: breakdownText,
-                    nonCacheInput: nonCachedTokens,
+                ratioType: ratioLabel,
-                    cacheInput: cacheTokens,
+                ratio: groupRatio,
-                    cacheRatio: cacheRatio,
+                symbol: symbol,
-                    cacheCreationInput: cacheCreationTokens,
+                total: (price * rate).toFixed(6),
-                    cacheCreationRatio: cacheCreationRatio,
+              },
-                    symbol: symbol,
+            )}
-                    cachePrice: (cacheRatioPrice * rate).toFixed(2),
-                    cacheCreationPrice: (
-                      cacheCreationRatioPrice * rate
-                    ).toFixed(6),
-                    price: (inputRatioPrice * rate).toFixed(6),
-                    completion: completionTokens,
-                    compPrice: (completionRatioPrice * rate).toFixed(6),
-                    ratio: groupRatio,
-                    ratioType: ratioLabel,
-                    total: (price * rate).toFixed(6),
-                  },
-                )
-              : i18next.t(
-                  '提示 {{input}} tokens / 1M tokens * {{symbol}}{{price}} + 补全 {{completion}} tokens / 1M tokens * {{symbol}}{{compPrice}} * {{ratioType}} {{ratio}} = {{symbol}}{{total}}',
-                  {
-                    input: inputTokens,
-                    symbol: symbol,
-                    price: (inputRatioPrice * rate).toFixed(6),
-                    completion: completionTokens,
-                    compPrice: (completionRatioPrice * rate).toFixed(6),
-                    ratio: groupRatio,
-                    ratioType: ratioLabel,
-                    total: (price * rate).toFixed(6),
-                  },
-                )}
           </p>
           <p>{i18next.t('仅供参考，以实际扣费为准')}</p>
         </article>
@@ -1825,6 +1983,10 @@ export function renderClaudeLogContent(
   user_group_ratio,
   cacheRatio = 1.0,
   cacheCreationRatio = 1.0,
+  cacheCreationTokens5m = 0,
+  cacheCreationRatio5m = 1.0,
+  cacheCreationTokens1h = 0,
+  cacheCreationRatio1h = 1.0,
 ) {
   const { ratio: effectiveGroupRatio, label: ratioLabel } = getEffectiveRatio(
     groupRatio,
@@ -1843,17 +2005,58 @@ export function renderClaudeLogContent(
       ratio: groupRatio,
     });
   } else {
-    return i18next.t(
+    const hasSplitCacheCreation =
-      '模型倍率 {{modelRatio}}，输出倍率 {{completionRatio}}，缓存倍率 {{cacheRatio}}，缓存创建倍率 {{cacheCreationRatio}}，{{ratioType}} {{ratio}}',
+      cacheCreationTokens5m > 0 || cacheCreationTokens1h > 0;
-      {
+    const shouldShowCacheCreation5m =
-        modelRatio: modelRatio,
+      hasSplitCacheCreation && cacheCreationTokens5m > 0;
-        completionRatio: completionRatio,
+    const shouldShowCacheCreation1h =
-        cacheRatio: cacheRatio,
+      hasSplitCacheCreation && cacheCreationTokens1h > 0;
-        cacheCreationRatio: cacheCreationRatio,
+
+    let cacheCreationPart = null;
+    if (hasSplitCacheCreation) {
+      if (shouldShowCacheCreation5m && shouldShowCacheCreation1h) {
+        cacheCreationPart = i18next.t(
+          '缓存创建倍率 5m {{cacheCreationRatio5m}} / 1h {{cacheCreationRatio1h}}',
+          {
+            cacheCreationRatio5m,
+            cacheCreationRatio1h,
+          },
+        );
+      } else if (shouldShowCacheCreation5m) {
+        cacheCreationPart = i18next.t(
+          '缓存创建倍率 5m {{cacheCreationRatio5m}}',
+          {
+            cacheCreationRatio5m,
+          },
+        );
+      } else if (shouldShowCacheCreation1h) {
+        cacheCreationPart = i18next.t(
+          '缓存创建倍率 1h {{cacheCreationRatio1h}}',
+          {
+            cacheCreationRatio1h,
+          },
+        );
+      }
+    }
+
+    if (!cacheCreationPart) {
+      cacheCreationPart = i18next.t('缓存创建倍率 {{cacheCreationRatio}}', {
+        cacheCreationRatio,
+      });
+    }
+
+    const parts = [
+      i18next.t('模型倍率 {{modelRatio}}', { modelRatio }),
+      i18next.t('输出倍率 {{completionRatio}}', { completionRatio }),
+      i18next.t('缓存倍率 {{cacheRatio}}', { cacheRatio }),
+      cacheCreationPart,
+      i18next.t('{{ratioType}} {{ratio}}', {
         ratioType: ratioLabel,
         ratio: groupRatio,
-      },
+      }),
-    );
+    ];
+
+    return parts.join('，');
   }
 }
 

web/src/hooks/tokens/useTokensData.jsx

@@ -20,7 +20,13 @@ For commercial licensing, please contact support@quantumnous.com
 import { useState, useEffect } from 'react';
 import { useTranslation } from 'react-i18next';
 import { Modal } from '@douyinfe/semi-ui';
-import { API, copy, showError, showSuccess } from '../../helpers';
+import {
+  API,
+  copy,
+  showError,
+  showSuccess,
+  encodeToBase64,
+} from '../../helpers';
 import { ITEMS_PER_PAGE } from '../../constants';
 import { useTableCompactMode } from '../common/useTableCompactMode';
 
@@ -136,7 +142,7 @@ export const useTokensData = (openFluentNotification) => {
         apiKey: 'sk-' + record.key,
       };
       let encodedConfig = encodeURIComponent(
-        btoa(JSON.stringify(cherryConfig)),
+        encodeToBase64(JSON.stringify(cherryConfig)),
       );
       url = url.replaceAll('{cherryConfig}', encodedConfig);
     } else {

web/src/hooks/usage-logs/useUsageLogsData.jsx

@@ -361,6 +361,10 @@ export const useLogsData = () => {
                 other?.user_group_ratio,
                 other.cache_ratio || 1.0,
                 other.cache_creation_ratio || 1.0,
+                other.cache_creation_tokens_5m || 0,
+                other.cache_creation_ratio_5m || other.cache_creation_ratio || 1.0,
+                other.cache_creation_tokens_1h || 0,
+                other.cache_creation_ratio_1h || other.cache_creation_ratio || 1.0,
               )
             : renderLogContent(
                 other?.model_ratio,
@@ -429,6 +433,10 @@ export const useLogsData = () => {
             other.cache_ratio || 1.0,
             other.cache_creation_tokens || 0,
             other.cache_creation_ratio || 1.0,
+            other.cache_creation_tokens_5m || 0,
+            other.cache_creation_ratio_5m || other.cache_creation_ratio || 1.0,
+            other.cache_creation_tokens_1h || 0,
+            other.cache_creation_ratio_1h || other.cache_creation_ratio || 1.0,
           );
         } else {
           content = renderModelPrice(

web/src/i18n/locales/en.json

@@ -1516,6 +1516,10 @@
     "缓存倍率": "Cache ratio",
     "缓存创建 Tokens": "Cache Creation Tokens",
     "缓存创建: {{cacheCreationRatio}}": "Cache creation: {{cacheCreationRatio}}",
+    "缓存创建: 5m {{cacheCreationRatio5m}}": "Cache creation: 5m {{cacheCreationRatio5m}}",
+    "缓存创建: 1h {{cacheCreationRatio1h}}": "Cache creation: 1h {{cacheCreationRatio1h}}",
+    "缓存创建倍率 5m {{cacheCreationRatio5m}}": "Cache creation multiplier 5m {{cacheCreationRatio5m}}",
+    "缓存创建倍率 1h {{cacheCreationRatio1h}}": "Cache creation multiplier 1h {{cacheCreationRatio1h}}",
     "缓存创建价格：{{symbol}}{{price}} * {{ratio}} = {{symbol}}{{total}} / 1M tokens (缓存创建倍率: {{cacheCreationRatio}})": "Cache creation price: {{symbol}}{{price}} * {{ratio}} = {{symbol}}{{total}} / 1M tokens (Cache creation ratio: {{cacheCreationRatio}})",
     "编辑": "Edit",
     "编辑API": "Edit API",
@@ -2104,4 +2108,4 @@
     "统一的": "The Unified",
     "大模型接口网关": "LLM API Gateway"
   }
-}
+}

web/src/i18n/locales/fr.json

@@ -1525,6 +1525,10 @@
     "缓存倍率": "Ratio de cache",
     "缓存创建 Tokens": "Jetons de création de cache",
     "缓存创建: {{cacheCreationRatio}}": "Création de cache : {{cacheCreationRatio}}",
+    "缓存创建: 5m {{cacheCreationRatio5m}}": "Création de cache : 5m {{cacheCreationRatio5m}}",
+    "缓存创建: 1h {{cacheCreationRatio1h}}": "Création de cache : 1h {{cacheCreationRatio1h}}",
+    "缓存创建倍率 5m {{cacheCreationRatio5m}}": "Multiplicateur de création de cache 5m {{cacheCreationRatio5m}}",
+    "缓存创建倍率 1h {{cacheCreationRatio1h}}": "Multiplicateur de création de cache 1h {{cacheCreationRatio1h}}",
     "缓存创建价格：{{symbol}}{{price}} * {{ratio}} = {{symbol}}{{total}} / 1M tokens (缓存创建倍率: {{cacheCreationRatio}})": "Prix de création du cache : {{symbol}}{{price}} * {{ratio}} = {{symbol}}{{total}} / 1M tokens (taux de création de cache : {{cacheCreationRatio}})",
     "编辑": "Modifier",
     "编辑API": "Modifier l'API",

web/src/i18n/locales/ja.json

@@ -1516,6 +1516,10 @@
     "缓存倍率": "キャッシュ倍率",
     "缓存创建 Tokens": "キャッシュ作成トークン",
     "缓存创建: {{cacheCreationRatio}}": "キャッシュ作成：{{cacheCreationRatio}}",
+    "缓存创建: 5m {{cacheCreationRatio5m}}": "キャッシュ作成：5m {{cacheCreationRatio5m}}",
+    "缓存创建: 1h {{cacheCreationRatio1h}}": "キャッシュ作成：1h {{cacheCreationRatio1h}}",
+    "缓存创建倍率 5m {{cacheCreationRatio5m}}": "キャッシュ作成倍率 5m {{cacheCreationRatio5m}}",
+    "缓存创建倍率 1h {{cacheCreationRatio1h}}": "キャッシュ作成倍率 1h {{cacheCreationRatio1h}}",
     "缓存创建价格：{{symbol}}{{price}} * {{ratio}} = {{symbol}}{{total}} / 1M tokens (缓存创建倍率: {{cacheCreationRatio}})": "キャッシュ作成料金：{{symbol}}{{price}} * {{ratio}} = {{symbol}}{{total}} / 1Mtokens（キャッシュ作成倍率：{{cacheCreationRatio}}）",
     "编辑": "編集",
     "编辑API": "API編集",
@@ -2075,4 +2079,4 @@
     "统一的": "統合型",
     "大模型接口网关": "LLM APIゲートウェイ"
   }
-}
+}

web/src/i18n/locales/ru.json

@@ -1534,6 +1534,10 @@
     "缓存倍率": "Коэффициент кэширования",
     "缓存创建 Tokens": "Создание кэша токенов",
     "缓存创建: {{cacheCreationRatio}}": "Создание кэша: {{cacheCreationRatio}}",
+    "缓存创建: 5m {{cacheCreationRatio5m}}": "Создание кэша: 5m {{cacheCreationRatio5m}}",
+    "缓存创建: 1h {{cacheCreationRatio1h}}": "Создание кэша: 1h {{cacheCreationRatio1h}}",
+    "缓存创建倍率 5m {{cacheCreationRatio5m}}": "Множитель создания кэша 5m {{cacheCreationRatio5m}}",
+    "缓存创建倍率 1h {{cacheCreationRatio1h}}": "Множитель создания кэша 1h {{cacheCreationRatio1h}}",
     "缓存创建价格：{{symbol}}{{price}} * {{ratio}} = {{symbol}}{{total}} / 1M tokens (缓存创建倍率: {{cacheCreationRatio}})": "Цена создания кэша: {{symbol}}{{price}} * {{ratio}} = {{symbol}}{{total}} / 1M токенов (коэффициент создания кэша: {{cacheCreationRatio}})",
     "编辑": "Редактировать",
     "编辑API": "Редактировать API",

web/src/i18n/locales/zh.json

@@ -1507,6 +1507,10 @@
     "缓存倍率": "缓存倍率",
     "缓存创建 Tokens": "缓存创建 Tokens",
     "缓存创建: {{cacheCreationRatio}}": "缓存创建: {{cacheCreationRatio}}",
+    "缓存创建: 5m {{cacheCreationRatio5m}}": "缓存创建: 5m {{cacheCreationRatio5m}}",
+    "缓存创建: 1h {{cacheCreationRatio1h}}": "缓存创建: 1h {{cacheCreationRatio1h}}",
+    "缓存创建倍率 5m {{cacheCreationRatio5m}}": "缓存创建倍率 5m {{cacheCreationRatio5m}}",
+    "缓存创建倍率 1h {{cacheCreationRatio1h}}": "缓存创建倍率 1h {{cacheCreationRatio1h}}",
     "缓存创建价格：{{symbol}}{{price}} * {{ratio}} = {{symbol}}{{total}} / 1M tokens (缓存创建倍率: {{cacheCreationRatio}})": "缓存创建价格：{{symbol}}{{price}} * {{ratio}} = {{symbol}}{{total}} / 1M tokens (缓存创建倍率: {{cacheCreationRatio}})",
     "编辑": "编辑",
     "编辑API": "编辑API",
@@ -2066,4 +2070,4 @@
     "Creem 介绍": "Creem 是一个简单的支付处理平台，支持固定金额产品销售，以及订阅销售。",
     "Creem Setting Tips": "Creem 只支持预设的固定金额产品，这产品以及价格需要提前在Creem网站内创建配置，所以不支持自定义动态金额充值。在Creem端配置产品的名字以及价格，获取Product Id 后填到下面的产品，在new-api为该产品设置充值额度，以及展示价格。"
   }
-}
+}

web/src/pages/Playground/index.jsx

@@ -47,6 +47,7 @@ import {
   createLoadingAssistantMessage,
   getTextContent,
   buildApiPayload,
+  encodeToBase64,
 } from '../../helpers';
 
 // Components
@@ -72,7 +73,7 @@ const generateAvatarDataUrl = (username) => {
       <text x="50%" y="50%" dominant-baseline="central" text-anchor="middle" font-size="16" fill="#ffffff" font-family="sans-serif">${firstLetter}</text>
     </svg>
   `;
-  return `data:image/svg+xml;base64,${btoa(svg)}`;
+  return `data:image/svg+xml;base64,${encodeToBase64(svg)}`;
 };
 
 const Playground = () => {