feat(channel): add per-channel billing ratio (渠道倍率)

constant/context_key.go

@@ -21,6 +21,7 @@ const (
 	ContextKeyTokenCrossGroupRetry   ContextKey = "token_cross_group_retry"
 
 	/* channel related keys */
+	ContextKeyChannelRatio             ContextKey = "channel_ratio"
 	ContextKeyChannelId                ContextKey = "channel_id"
 	ContextKeyChannelName              ContextKey = "channel_name"
 	ContextKeyChannelCreateTime        ContextKey = "channel_create_time"

controller/relay.go

@@ -584,6 +584,7 @@ func RelayTask(c *gin.Context) {
 		task.PrivateData.BillingContext = &model.TaskBillingContext{
 			ModelPrice:      relayInfo.PriceData.ModelPrice,
 			GroupRatio:      relayInfo.PriceData.GroupRatioInfo.GroupRatio,
+			ChannelRatio:    relayInfo.PriceData.ChannelRatio,
 			ModelRatio:      relayInfo.PriceData.ModelRatio,
 			OtherRatios:     relayInfo.PriceData.OtherRatios,
 			OriginModelName: relayInfo.OriginModelName,

middleware/distributor.go

@@ -485,6 +485,7 @@ func SetupContextForSelectedChannel(c *gin.Context, channel *model.Channel, mode
 	case constant.ChannelTypeCoze:
 		c.Set("bot_id", channel.Other)
 	}
+	common.SetContextKey(c, constant.ContextKeyChannelRatio, channel.GetRatio())
 	return nil
 }
 

model/ability.go

@@ -25,13 +25,15 @@ type Ability struct {
 
 type AbilityWithChannel struct {
 	Ability
-	ChannelType int `json:"channel_type"`
+	ChannelType   int     `json:"channel_type"`
+	ChannelRatio  float64 `json:"channel_ratio"`
+	ChannelWeight uint    `json:"channel_weight"`
 }
 
 func GetAllEnableAbilityWithChannels() ([]AbilityWithChannel, error) {
 	var abilities []AbilityWithChannel
 	err := DB.Table("abilities").
-		Select("abilities.*, channels.type as channel_type").
+		Select("abilities.*, channels.type as channel_type, COALESCE(channels.ratio, 1) as channel_ratio, COALESCE(channels.weight, 0) as channel_weight").
 		Joins("left join channels on abilities.channel_id = channels.id").
 		Where("abilities.enabled = ?", true).
 		Scan(&abilities).Error

model/channel.go

@@ -55,6 +55,10 @@ type Channel struct {
 
 	OtherSettings string `json:"settings" gorm:"column:settings"` // 其他设置，存储azure版本等不需要检索的信息，详见dto.ChannelOtherSettings
 
+	// Ratio is a per-channel billing multiplier applied on top of model and group ratios.
+	// nil or 0 is treated as 1.0 (no adjustment). Values < 1 reduce cost, > 1 increase cost.
+	Ratio *float64 `json:"ratio" gorm:"default:1"`
+
 	// cache info
 	Keys []string `json:"-" gorm:"-"`
 }
@@ -342,6 +346,13 @@ func (channel *Channel) GetAutoBan() bool {
 	return *channel.AutoBan == 1
 }
 
+func (channel *Channel) GetRatio() float64 {
+	if channel.Ratio == nil || *channel.Ratio == 0 {
+		return 1.0
+	}
+	return *channel.Ratio
+}
+
 func (channel *Channel) Save() error {
 	return DB.Save(channel).Error
 }

model/pricing.go

@@ -36,6 +36,12 @@ type Pricing struct {
 	BillingMode            string                  `json:"billing_mode,omitempty"`
 	BillingExpr            string                  `json:"billing_expr,omitempty"`
 	PricingVersion         string                  `json:"pricing_version,omitempty"`
+	// GroupChannelRatioMin holds the minimum channel ratio per group across all channels serving that model+group.
+	// Omitted when all channels use the default ratio of 1.0.
+	GroupChannelRatioMin map[string]float64 `json:"group_channel_ratio_min,omitempty"`
+	// GroupChannelRatioMax holds the maximum channel ratio per group.
+	// Omitted when equal to GroupChannelRatioMin (no range to show).
+	GroupChannelRatioMax map[string]float64 `json:"group_channel_ratio_max,omitempty"`
 }
 
 type PricingVendor struct {
@@ -190,13 +196,39 @@ func updatePricing() {
 
 	modelGroupsMap := make(map[string]*types.Set[string])
 
+	// modelGroupChannelRatio[model][group] = {min, max} channel ratio across all channels
+	type minMaxRatio struct {
+		min float64
+		max float64
+	}
+	modelGroupChannelRatio := make(map[string]map[string]*minMaxRatio)
+
 	for _, ability := range enableAbilities {
 		groups, ok := modelGroupsMap[ability.Model]
 		if !ok {
 			groups = types.NewSet[string]()
 			modelGroupsMap[ability.Model] = groups
 		}
 		groups.Add(ability.Group)
+
+		// accumulate min/max channel ratio per (model, group)
+		if _, ok := modelGroupChannelRatio[ability.Model]; !ok {
+			modelGroupChannelRatio[ability.Model] = make(map[string]*minMaxRatio)
+		}
+		cr := ability.ChannelRatio
+		if cr <= 0 {
+			cr = 1.0
+		}
+		if mm, exists := modelGroupChannelRatio[ability.Model][ability.Group]; exists {
+			if cr < mm.min {
+				mm.min = cr
+			}
+			if cr > mm.max {
+				mm.max = cr
+			}
+		} else {
+			modelGroupChannelRatio[ability.Model][ability.Group] = &minMaxRatio{min: cr, max: cr}
+		}
 	}
 
 	//这里使用切片而不是Set，因为一个模型可能支持多个端点类型，并且第一个端点是优先使用端点
@@ -337,6 +369,35 @@ func updatePricing() {
 				pricing.BillingExpr = expr
 			}
 		}
+
+		// compute per-group min/max channel ratio
+		if groupRatios, ok := modelGroupChannelRatio[model]; ok {
+			minMap := make(map[string]float64)
+			maxMap := make(map[string]float64)
+			for group, mm := range groupRatios {
+				if mm.min != 1.0 || mm.max != 1.0 {
+					minMap[group] = mm.min
+					maxMap[group] = mm.max
+				}
+			}
+			if len(minMap) > 0 {
+				pricing.GroupChannelRatioMin = minMap
+			}
+			if len(maxMap) > 0 {
+				// only store max when it differs from min somewhere
+				hasRange := false
+				for g, mx := range maxMap {
+					if mx != minMap[g] {
+						hasRange = true
+						break
+					}
+				}
+				if hasRange {
+					pricing.GroupChannelRatioMax = maxMap
+				}
+			}
+		}
+
 		pricingMap = append(pricingMap, pricing)
 	}
 

model/task.go

@@ -111,6 +111,7 @@ type TaskPrivateData struct {
 type TaskBillingContext struct {
 	ModelPrice      float64            `json:"model_price,omitempty"`       // 模型单价
 	GroupRatio      float64            `json:"group_ratio,omitempty"`       // 分组倍率
+	ChannelRatio    float64            `json:"channel_ratio,omitempty"`     // 渠道倍率
 	ModelRatio      float64            `json:"model_ratio,omitempty"`       // 模型倍率
 	OtherRatios     map[string]float64 `json:"other_ratios,omitempty"`      // 附加倍率（时长、分辨率等）
 	OriginModelName string             `json:"origin_model_name,omitempty"` // 模型名称，必须为OriginModelName

pkg/billingexpr/settle.go

@@ -23,7 +23,11 @@ func ComputeTieredQuotaWithRequest(snap *BillingSnapshot, params TokenParams, re
 	}
 
 	quotaBeforeGroup := quotaConversion(cost, snap)
-	afterGroup := QuotaRound(quotaBeforeGroup * snap.GroupRatio)
+	cr := snap.ChannelRatio
+	if cr <= 0 {
+		cr = 1.0
+	}
+	afterGroup := QuotaRound(quotaBeforeGroup * snap.GroupRatio * cr)
 	crossed := trace.MatchedTier != snap.EstimatedTier
 
 	return TieredResult{

pkg/billingexpr/types.go

@@ -42,6 +42,7 @@ type BillingSnapshot struct {
 	ExprString                string  `json:"expr_string"`
 	ExprHash                  string  `json:"expr_hash"`
 	GroupRatio                float64 `json:"group_ratio"`
+	ChannelRatio              float64 `json:"channel_ratio"`
 	EstimatedPromptTokens     int     `json:"estimated_prompt_tokens"`
 	EstimatedCompletionTokens int     `json:"estimated_completion_tokens"`
 	EstimatedQuotaBeforeGroup float64 `json:"estimated_quota_before_group"`

relay/common/relay_info.go

@@ -78,6 +78,7 @@ type ChannelMeta struct {
 	UpstreamModelName    string
 	IsModelMapped        bool
 	SupportStreamOptions bool // 是否支持流式选项
+	ChannelRatio         float64
 }
 
 type TokenCountMeta struct {
@@ -232,6 +233,12 @@ func (info *RelayInfo) InitChannelMeta(c *gin.Context) {
 		channelMeta.SupportStreamOptions = true
 	}
 
+	channelRatio, ok := common.GetContextKeyType[float64](c, constant.ContextKeyChannelRatio)
+	if !ok || channelRatio <= 0 {
+		channelRatio = 1.0
+	}
+	channelMeta.ChannelRatio = channelRatio
+
 	info.ChannelMeta = channelMeta
 
 	// reset some fields based on channel meta

relay/helper/price.go

@@ -5,6 +5,7 @@ import (
 	"strings"
 
 	"github.com/QuantumNous/new-api/common"
+	"github.com/QuantumNous/new-api/constant"
 	"github.com/QuantumNous/new-api/logger"
 	"github.com/QuantumNous/new-api/model"
 	"github.com/QuantumNous/new-api/pkg/billingexpr"
@@ -35,6 +36,14 @@ func modelPriceNotConfiguredError(modelName string, userId int) error {
 // https://docs.claude.com/en/docs/build-with-claude/prompt-caching#1-hour-cache-duration
 const claudeCacheCreation1hMultiplier = 6 / 3.75
 
+func getChannelRatioFromContext(c *gin.Context) float64 {
+	v, ok := common.GetContextKeyType[float64](c, constant.ContextKeyChannelRatio)
+	if !ok || v <= 0 {
+		return 1.0
+	}
+	return v
+}
+
 // 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{
@@ -69,9 +78,11 @@ func ModelPriceHelper(c *gin.Context, info *relaycommon.RelayInfo, promptTokens
 
 	groupRatioInfo := HandleGroupRatio(c, info)
 
+	channelRatio := getChannelRatioFromContext(c)
+
 	// Check if this model uses tiered_expr billing
 	if billing_setting.GetBillingMode(info.OriginModelName) == billing_setting.BillingModeTieredExpr {
-		return modelPriceHelperTiered(c, info, promptTokens, meta, groupRatioInfo)
+		return modelPriceHelperTiered(c, info, promptTokens, meta, groupRatioInfo, channelRatio)
 	}
 
 	var preConsumedQuota int
@@ -111,13 +122,13 @@ func ModelPriceHelper(c *gin.Context, info *relaycommon.RelayInfo, promptTokens
 		imageRatio, _ = ratio_setting.GetImageRatio(info.OriginModelName)
 		audioRatio = ratio_setting.GetAudioRatio(info.OriginModelName)
 		audioCompletionRatio = ratio_setting.GetAudioCompletionRatio(info.OriginModelName)
-		ratio := modelRatio * groupRatioInfo.GroupRatio
+		ratio := modelRatio * groupRatioInfo.GroupRatio * channelRatio
 		preConsumedQuota = int(float64(preConsumedTokens) * ratio)
 	} else {
 		if meta.ImagePriceRatio != 0 {
 			modelPrice = modelPrice * meta.ImagePriceRatio
 		}
-		preConsumedQuota = int(modelPrice * common.QuotaPerUnit * groupRatioInfo.GroupRatio)
+		preConsumedQuota = int(modelPrice * common.QuotaPerUnit * groupRatioInfo.GroupRatio * channelRatio)
 	}
 
 	// check if free model pre-consume is disabled
@@ -145,6 +156,7 @@ func ModelPriceHelper(c *gin.Context, info *relaycommon.RelayInfo, promptTokens
 		ModelRatio:           modelRatio,
 		CompletionRatio:      completionRatio,
 		GroupRatioInfo:       groupRatioInfo,
+		ChannelRatio:         channelRatio,
 		UsePrice:             usePrice,
 		CacheRatio:           cacheRatio,
 		ImageRatio:           imageRatio,
@@ -167,6 +179,8 @@ func ModelPriceHelper(c *gin.Context, info *relaycommon.RelayInfo, promptTokens
 func ModelPriceHelperPerCall(c *gin.Context, info *relaycommon.RelayInfo) (types.PriceData, error) {
 	groupRatioInfo := HandleGroupRatio(c, info)
 
+	channelRatio := getChannelRatioFromContext(c)
+
 	modelPrice, success := ratio_setting.GetModelPrice(info.OriginModelName, true)
 	usePrice := success
 	var modelRatio float64
@@ -194,7 +208,7 @@ func ModelPriceHelperPerCall(c *gin.Context, info *relaycommon.RelayInfo) (types
 	freeModel := false
 
 	if usePrice {
-		quota = int(modelPrice * common.QuotaPerUnit * groupRatioInfo.GroupRatio)
+		quota = int(modelPrice * common.QuotaPerUnit * groupRatioInfo.GroupRatio * channelRatio)
 		if !operation_setting.GetQuotaSetting().EnableFreeModelPreConsume {
 			if groupRatioInfo.GroupRatio == 0 || modelPrice == 0 {
 				quota = 0
@@ -203,7 +217,7 @@ func ModelPriceHelperPerCall(c *gin.Context, info *relaycommon.RelayInfo) (types
 		}
 	} else {
 		// 按量计费：以模型倍率的一半作为预扣额度
-		quota = int(modelRatio / 2 * common.QuotaPerUnit * groupRatioInfo.GroupRatio)
+		quota = int(modelRatio / 2 * common.QuotaPerUnit * groupRatioInfo.GroupRatio * channelRatio)
 		modelPrice = -1
 		if !operation_setting.GetQuotaSetting().EnableFreeModelPreConsume {
 			if groupRatioInfo.GroupRatio == 0 || modelRatio == 0 {
@@ -219,6 +233,7 @@ func ModelPriceHelperPerCall(c *gin.Context, info *relaycommon.RelayInfo) (types
 		ModelRatio:     modelRatio,
 		UsePrice:       usePrice,
 		Quota:          quota,
+		ChannelRatio:   channelRatio,
 		GroupRatioInfo: groupRatioInfo,
 	}
 	return priceData, nil
@@ -238,7 +253,7 @@ func HasModelBillingConfig(modelName string) bool {
 	return ok && strings.TrimSpace(expr) != ""
 }
 
-func modelPriceHelperTiered(c *gin.Context, info *relaycommon.RelayInfo, promptTokens int, meta *types.TokenCountMeta, groupRatioInfo types.GroupRatioInfo) (types.PriceData, error) {
+func modelPriceHelperTiered(c *gin.Context, info *relaycommon.RelayInfo, promptTokens int, meta *types.TokenCountMeta, groupRatioInfo types.GroupRatioInfo, channelRatio float64) (types.PriceData, error) {
 	exprStr, ok := billing_setting.GetBillingExpr(info.OriginModelName)
 	if !ok {
 		return types.PriceData{}, fmt.Errorf("model %s is configured as tiered_expr but has no billing expression", info.OriginModelName)
@@ -265,7 +280,7 @@ func modelPriceHelperTiered(c *gin.Context, info *relaycommon.RelayInfo, promptT
 
 	// Expression coefficients are $/1M tokens prices; convert to quota the same way per-call billing does.
 	quotaBeforeGroup := rawCost / 1_000_000 * common.QuotaPerUnit
-	preConsumedQuota := billingexpr.QuotaRound(quotaBeforeGroup * groupRatioInfo.GroupRatio)
+	preConsumedQuota := billingexpr.QuotaRound(quotaBeforeGroup * groupRatioInfo.GroupRatio * channelRatio)
 
 	freeModel := false
 	if !operation_setting.GetQuotaSetting().EnableFreeModelPreConsume {
@@ -282,6 +297,7 @@ func modelPriceHelperTiered(c *gin.Context, info *relaycommon.RelayInfo, promptT
 		ExprString:                exprStr,
 		ExprHash:                  exprHash,
 		GroupRatio:                groupRatioInfo.GroupRatio,
+		ChannelRatio:              channelRatio,
 		EstimatedPromptTokens:     promptTokens,
 		EstimatedCompletionTokens: estimatedCompletionTokens,
 		EstimatedQuotaBeforeGroup: quotaBeforeGroup,
@@ -296,6 +312,7 @@ func modelPriceHelperTiered(c *gin.Context, info *relaycommon.RelayInfo, promptT
 	priceData := types.PriceData{
 		FreeModel:         freeModel,
 		GroupRatioInfo:    groupRatioInfo,
+		ChannelRatio:      channelRatio,
 		QuotaToPreConsume: preConsumedQuota,
 	}
 

service/log_info_generate.go

@@ -43,6 +43,9 @@ func GenerateTextOtherInfo(ctx *gin.Context, relayInfo *relaycommon.RelayInfo, m
 	other["cache_ratio"] = cacheRatio
 	other["model_price"] = modelPrice
 	other["user_group_ratio"] = userGroupRatio
+	if channelRatio := relayInfo.PriceData.ChannelRatio; channelRatio != 0 && channelRatio != 1.0 {
+		other["channel_ratio"] = channelRatio
+	}
 	other["frt"] = float64(relayInfo.FirstResponseTime.UnixMilli() - relayInfo.StartTime.UnixMilli())
 	if relayInfo.ReasoningEffort != "" {
 		other["reasoning_effort"] = relayInfo.ReasoningEffort
@@ -261,6 +264,9 @@ func GenerateMjOtherInfo(relayInfo *relaycommon.RelayInfo, priceData types.Price
 	if priceData.GroupRatioInfo.HasSpecialRatio {
 		other["user_group_ratio"] = priceData.GroupRatioInfo.GroupSpecialRatio
 	}
+	if priceData.ChannelRatio != 0 && priceData.ChannelRatio != 1.0 {
+		other["channel_ratio"] = priceData.ChannelRatio
+	}
 	appendRequestPath(nil, relayInfo, other)
 	return other
 }