mirror of
https://github.com/Wei-Shaw/sub2api.git
synced 2026-10-07 14:38:27 +08:00
fix(pricing): align DeepSeek billing with official peak/off-peak rates
同步 DeepSeek 官方峰谷价格与模型定价。
This commit is contained in:
alfadb
@@ -11,6 +11,7 @@ import (
|
||||
"time"
|
||||
|
||||
"github.com/Wei-Shaw/sub2api/internal/config"
|
||||
"github.com/Wei-Shaw/sub2api/internal/pkg/timezone"
|
||||
"github.com/Wei-Shaw/sub2api/internal/pkg/xai"
|
||||
)
|
||||
|
||||
@@ -230,6 +231,46 @@ func resolvedChannelTimeMultiplier(resolved *ResolvedPricing, at time.Time) floa
|
||||
// sources can price the requested model.
|
||||
var ErrModelPricingUnavailable = errors.New("pricing not found")
|
||||
|
||||
// ---- DeepSeek 官方低谷价（$/token，2026-08-23 起生效）----
|
||||
// Source: https://api-docs.deepseek.com/quick_start/pricing
|
||||
// 高峰价 = 2× 低谷价；高峰时段 01:00–04:00 与 06:00–10:00 UTC（仅工作日），
|
||||
// 北京时间周六/周日全天低谷。时段判定见 deepseekPeakMultiplierAt。
|
||||
const (
|
||||
deepseekFlashOffPeakInputPrice = 2.2e-7 // $0.22 per MTok (cache miss)
|
||||
deepseekFlashOffPeakOutputPrice = 6.6e-7 // $0.66 per MTok
|
||||
deepseekFlashOffPeakCacheRead = 7e-9 // $0.007 per MTok (cache hit)
|
||||
deepseekProOffPeakInputPrice = 6.6e-7 // $0.66 per MTok (cache miss)
|
||||
deepseekProOffPeakOutputPrice = 1.98e-6 // $1.98 per MTok
|
||||
deepseekProOffPeakCacheRead = 2.2e-8 // $0.022 per MTok (cache hit)
|
||||
)
|
||||
|
||||
// isDeepSeekModel 判断模型名是否为 DeepSeek 模型（大小写不敏感）。
|
||||
// 任意 deepseek- 前缀均视为 DeepSeek 模型：官方模型（v4-flash / v4-pro /
|
||||
// v4-flash-vision-exp）按各自价卡计价，其余 deepseek-*（含已停服的
|
||||
// deepseek-chat / deepseek-reasoner 与未知型号）统一按 flash 价兜底，
|
||||
// 避免计费中断；新名字由 fallback warn 日志（每模型每进程一条）暴露，
|
||||
// 运营者据此更新价卡。
|
||||
func isDeepSeekModel(model string) bool {
|
||||
return strings.HasPrefix(strings.ToLower(strings.TrimSpace(model)), "deepseek-")
|
||||
}
|
||||
|
||||
// deepseekPeakMultiplierAt 返回指定时刻的 DeepSeek 官方峰谷定价因子。
|
||||
// 官方口径（2026-08-23 起生效）：高峰价 = 2× 低谷价；高峰时段为
|
||||
// 01:00–04:00 与 06:00–10:00 UTC（半开区间），仅工作日；
|
||||
// 周末（北京时间周六/周日）全天低谷。北京时间用固定 +8 偏移（无夏令时）。
|
||||
func deepseekPeakMultiplierAt(now time.Time) float64 {
|
||||
beijing := now.In(time.FixedZone("Asia/Shanghai", 8*3600))
|
||||
switch beijing.Weekday() {
|
||||
case time.Saturday, time.Sunday:
|
||||
return 1.0
|
||||
}
|
||||
switch h := now.UTC().Hour(); {
|
||||
case h >= 1 && h < 4, h >= 6 && h < 10:
|
||||
return 2.0
|
||||
}
|
||||
return 1.0
|
||||
}
|
||||
|
||||
// BillingService 计费服务
|
||||
type BillingService struct {
|
||||
cfg *config.Config
|
||||
@@ -466,19 +507,30 @@ func (s *BillingService) initFallbackPricing() {
|
||||
// 覆盖逻辑见同文件 getFallbackPricing()
|
||||
// ============================================================
|
||||
|
||||
// ---- DeepSeek V4 系列 ----
|
||||
// ---- DeepSeek 系列 ----
|
||||
// Source: https://api-docs.deepseek.com/quick_start/pricing
|
||||
// （deepseek-chat / deepseek-reasoner 为 deepseek-v4-flash 的兼容别名，2026/07/24 弃用）
|
||||
// 官方口径（2026-08-23 起生效）：现行模型为 deepseek-v4-flash /
|
||||
// deepseek-v4-pro / deepseek-v4-flash-vision-exp；deepseek-chat /
|
||||
// deepseek-reasoner 已停止服务，其余 deepseek-*（含未知型号）统一按
|
||||
// flash 价兜底（见 getFallbackPricing），避免计费中断。
|
||||
// 以下均为官方低谷价；高峰价 = 2× 低谷价（高峰时段 01:00–04:00
|
||||
// 与 06:00–10:00 UTC，仅工作日；北京时间周六/周日全天低谷），见 deepseekPeakMultiplierAt。
|
||||
s.fallbackPrices["deepseek-v4-pro"] = &ModelPricing{
|
||||
InputPricePerToken: 4.35e-7, // $0.435 per MTok (cache miss)
|
||||
OutputPricePerToken: 8.7e-7, // $0.87 per MTok
|
||||
CacheReadPricePerToken: 3.625e-9, // $0.003625 per MTok (cache hit)
|
||||
InputPricePerToken: deepseekProOffPeakInputPrice, // $0.66 per MTok (cache miss, off-peak)
|
||||
OutputPricePerToken: deepseekProOffPeakOutputPrice, // $1.98 per MTok
|
||||
CacheReadPricePerToken: deepseekProOffPeakCacheRead, // $0.022 per MTok (cache hit)
|
||||
SupportsCacheBreakdown: false,
|
||||
}
|
||||
s.fallbackPrices["deepseek-v4-flash"] = &ModelPricing{
|
||||
InputPricePerToken: 1.4e-7, // $0.14 per MTok (cache miss)
|
||||
OutputPricePerToken: 2.8e-7, // $0.28 per MTok
|
||||
CacheReadPricePerToken: 2.8e-9, // $0.0028 per MTok (cache hit)
|
||||
InputPricePerToken: deepseekFlashOffPeakInputPrice, // $0.22 per MTok (cache miss, off-peak)
|
||||
OutputPricePerToken: deepseekFlashOffPeakOutputPrice, // $0.66 per MTok
|
||||
CacheReadPricePerToken: deepseekFlashOffPeakCacheRead, // $0.007 per MTok (cache hit)
|
||||
SupportsCacheBreakdown: false,
|
||||
}
|
||||
s.fallbackPrices["deepseek-v4-flash-vision-exp"] = &ModelPricing{
|
||||
InputPricePerToken: deepseekFlashOffPeakInputPrice,
|
||||
OutputPricePerToken: deepseekFlashOffPeakOutputPrice,
|
||||
CacheReadPricePerToken: deepseekFlashOffPeakCacheRead,
|
||||
SupportsCacheBreakdown: false,
|
||||
}
|
||||
|
||||
@@ -796,15 +848,21 @@ func (s *BillingService) getFallbackPricing(model string) *ModelPricing {
|
||||
return s.fallbackPrices["gemini-3.6-flash"]
|
||||
}
|
||||
|
||||
// DeepSeek V4 系列：仅匹配已知 V4 Pro/Flash 与官方兼容别名
|
||||
// （deepseek-chat / deepseek-reasoner → V4 Flash），未知 deepseek-* 型号不回退，避免误计价。
|
||||
// DeepSeek 系列：官方模型 V4 Pro/Flash（含 vision-exp）按各自价卡；
|
||||
// 其余 deepseek-*（含已停服的 deepseek-chat / deepseek-reasoner 与未知型号）
|
||||
// 统一按 flash 价兜底，避免计费中断。新名字由 fallback warn 日志
|
||||
// （每模型每进程一条）暴露，运营者据此更新价卡。
|
||||
// "deepseek-v4-flash-vision-exp" 含 "deepseek-v4-flash" 子串，显式分支置于 flash 之前，语义清晰。
|
||||
if strings.Contains(modelLower, "deepseek-v4-flash-vision-exp") {
|
||||
return s.fallbackPrices["deepseek-v4-flash-vision-exp"]
|
||||
}
|
||||
if strings.Contains(modelLower, "deepseek-v4-flash") {
|
||||
return s.fallbackPrices["deepseek-v4-flash"]
|
||||
}
|
||||
if strings.Contains(modelLower, "deepseek-v4-pro") {
|
||||
return s.fallbackPrices["deepseek-v4-pro"]
|
||||
}
|
||||
if strings.Contains(modelLower, "deepseek-chat") || strings.Contains(modelLower, "deepseek-reasoner") {
|
||||
if strings.HasPrefix(modelLower, "deepseek-") {
|
||||
return s.fallbackPrices["deepseek-v4-flash"]
|
||||
}
|
||||
|
||||
@@ -1246,7 +1304,28 @@ func (s *BillingService) calculateTokenCost(resolved *ResolvedPricing, input Cos
|
||||
return nil, fmt.Errorf("no pricing available for model: %s: %w", input.Model, ErrModelPricingUnavailable)
|
||||
}
|
||||
|
||||
pricing = s.applyModelSpecificPricingPolicy(input.Model, pricing)
|
||||
// 默认价卡（Source=LiteLLM）应用 DeepSeek 官方价强制覆盖（幂等，GetModelPricing
|
||||
// 内部已强制过）；分组/渠道自定义定价保留运营者配置，不强制覆盖官方价。
|
||||
pricing = s.applyModelSpecificPricingPolicyEx(input.Model, pricing, resolved.Source == PricingSourceLiteLLM)
|
||||
|
||||
// DeepSeek 模型默认价卡按官方峰谷口径调整：高峰时段（01:00–04:00 与
|
||||
// 06:00–10:00 UTC，仅工作日；北京时间周末全天低谷）按 2× 低谷价计费。
|
||||
// 仅作用于默认价卡（Source=LiteLLM，无分组/渠道自定义定价）——分组/渠道
|
||||
// 自定义定价保持运营者语义，不叠加。PricingAt 为零值时回退当前时刻。
|
||||
// 先克隆再乘，避免污染共享 fallbackPrices 指针。
|
||||
if resolved.Source == PricingSourceLiteLLM && isDeepSeekModel(input.Model) {
|
||||
pricingAt := input.PricingAt
|
||||
if pricingAt.IsZero() {
|
||||
pricingAt = timezone.Now()
|
||||
}
|
||||
if mult := deepseekPeakMultiplierAt(pricingAt); mult > 1 {
|
||||
cloned := *pricing
|
||||
cloned.InputPricePerToken *= mult
|
||||
cloned.OutputPricePerToken *= mult
|
||||
cloned.CacheReadPricePerToken *= mult
|
||||
pricing = &cloned
|
||||
}
|
||||
}
|
||||
|
||||
// 官方长上下文阶梯仅在无区间定价时应用（区间定价已包含上下文分层）。
|
||||
applyLongCtx := len(resolved.Intervals) == 0 && contextTierPricingEnabled
|
||||
@@ -1489,10 +1568,43 @@ func (s *BillingService) calculateCostInternalWithPolicy(
|
||||
return s.computeTokenBreakdown(pricing, tokens, rateMultiplier, serviceTier, longContextBillingEnabled), nil
|
||||
}
|
||||
|
||||
// applyModelSpecificPricingPolicy 应用模型特定定价策略（GPT-5.6 长上下文/缓存写入、
|
||||
// DeepSeek 官方价强制覆盖等）。默认强制 DeepSeek 官方价——该路径仅被默认价卡
|
||||
// （GetModelPricing 内部）调用；分组/渠道自定义定价路径用带参数的
|
||||
// applyModelSpecificPricingPolicyEx 关闭强制，保留运营者配置。
|
||||
func (s *BillingService) applyModelSpecificPricingPolicy(model string, pricing *ModelPricing) *ModelPricing {
|
||||
return s.applyModelSpecificPricingPolicyEx(model, pricing, true)
|
||||
}
|
||||
|
||||
// applyModelSpecificPricingPolicyEx 与 applyModelSpecificPricingPolicy 相同，
|
||||
// 但由调用方控制是否强制 DeepSeek 官方价（forceDeepSeekRates）。
|
||||
// calculateTokenCost 对分组/渠道自定义定价（Source 非 LiteLLM）传 false：
|
||||
// 强制覆盖会把运营者配置的售价盖回官方价，违反自定义定价语义。
|
||||
func (s *BillingService) applyModelSpecificPricingPolicyEx(model string, pricing *ModelPricing, forceDeepSeekRates bool) *ModelPricing {
|
||||
if pricing == nil {
|
||||
return nil
|
||||
}
|
||||
// DeepSeek 模型：无论 JSON/远端价格表给什么价，一律强制官方低谷价
|
||||
// （2026-08-23 起生效）。这是覆盖远端旧价的关键——远端仓库不可改，生产会先
|
||||
// 拉到旧价，必须在此兜底修正；克隆后再覆盖，避免污染共享 fallbackPrices 指针。
|
||||
// 档位判定：含 "deepseek-v4-pro" 的版本化名称（如 deepseek-v4-pro-0813）归 pro 档，
|
||||
// 其余 deepseek-*（含已停服的 chat/reasoner 与未知型号）统一归 flash 档。
|
||||
// 高峰时段倍率不在本函数处理，由 calculateTokenCost 按 deepseekPeakMultiplierAt
|
||||
// 对默认价卡另行叠加（分组/渠道自定义定价不叠加）。
|
||||
if forceDeepSeekRates && isDeepSeekModel(model) {
|
||||
cloned := *pricing
|
||||
if strings.Contains(strings.ToLower(strings.TrimSpace(model)), "deepseek-v4-pro") {
|
||||
cloned.InputPricePerToken = deepseekProOffPeakInputPrice
|
||||
cloned.OutputPricePerToken = deepseekProOffPeakOutputPrice
|
||||
cloned.CacheReadPricePerToken = deepseekProOffPeakCacheRead
|
||||
} else {
|
||||
// deepseek-v4-flash / deepseek-v4-flash-vision-exp 与其余 deepseek-* 共用 flash 价。
|
||||
cloned.InputPricePerToken = deepseekFlashOffPeakInputPrice
|
||||
cloned.OutputPricePerToken = deepseekFlashOffPeakOutputPrice
|
||||
cloned.CacheReadPricePerToken = deepseekFlashOffPeakCacheRead
|
||||
}
|
||||
return &cloned
|
||||
}
|
||||
normalized := normalizeKnownOpenAICodexModel(model)
|
||||
isGPT56 := isOpenAIGPT56Model(normalized)
|
||||
usesLegacyLongContextPricing := usesOpenAILegacyLongContextPricing(normalized)
|
||||
|
||||
@@ -490,30 +490,45 @@ func TestGetFallbackPricing_FamilyMatching(t *testing.T) {
|
||||
{
|
||||
name: "deepseek v4 pro",
|
||||
model: "deepseek-v4-pro",
|
||||
expectedInput: 4.35e-7,
|
||||
expectedOutput: floatPtr(8.7e-7),
|
||||
expectedCacheRead: floatPtr(3.625e-9),
|
||||
expectedInput: 6.6e-7,
|
||||
expectedOutput: floatPtr(1.98e-6),
|
||||
expectedCacheRead: floatPtr(2.2e-8),
|
||||
},
|
||||
{
|
||||
name: "deepseek v4 flash",
|
||||
model: "deepseek-v4-flash",
|
||||
expectedInput: 1.4e-7,
|
||||
expectedOutput: floatPtr(2.8e-7),
|
||||
expectedCacheRead: floatPtr(2.8e-9),
|
||||
expectedInput: 2.2e-7,
|
||||
expectedOutput: floatPtr(6.6e-7),
|
||||
expectedCacheRead: floatPtr(7e-9),
|
||||
},
|
||||
{
|
||||
name: "deepseek chat alias → flash",
|
||||
name: "deepseek v4 flash vision exp",
|
||||
model: "deepseek-v4-flash-vision-exp",
|
||||
expectedInput: 2.2e-7,
|
||||
expectedOutput: floatPtr(6.6e-7),
|
||||
expectedCacheRead: floatPtr(7e-9),
|
||||
},
|
||||
{
|
||||
// deepseek-chat / deepseek-reasoner 已停止服务，统一按 flash 价兜底。
|
||||
name: "deepseek chat discontinued maps to flash",
|
||||
model: "deepseek-chat",
|
||||
expectedInput: 1.4e-7,
|
||||
expectedOutput: floatPtr(2.8e-7),
|
||||
expectedCacheRead: floatPtr(2.8e-9),
|
||||
expectedInput: 2.2e-7,
|
||||
expectedOutput: floatPtr(6.6e-7),
|
||||
expectedCacheRead: floatPtr(7e-9),
|
||||
},
|
||||
{
|
||||
name: "deepseek reasoner alias → flash",
|
||||
name: "deepseek reasoner discontinued maps to flash",
|
||||
model: "deepseek-reasoner",
|
||||
expectedInput: 1.4e-7,
|
||||
expectedOutput: floatPtr(2.8e-7),
|
||||
expectedCacheRead: floatPtr(2.8e-9),
|
||||
expectedInput: 2.2e-7,
|
||||
expectedOutput: floatPtr(6.6e-7),
|
||||
expectedCacheRead: floatPtr(7e-9),
|
||||
},
|
||||
{
|
||||
name: "unknown deepseek maps to flash",
|
||||
model: "deepseek-foo",
|
||||
expectedInput: 2.2e-7,
|
||||
expectedOutput: floatPtr(6.6e-7),
|
||||
expectedCacheRead: floatPtr(7e-9),
|
||||
},
|
||||
|
||||
// ---- 智谱 GLM（z.ai USD 口径）----
|
||||
|
||||
@@ -0,0 +1,345 @@
|
||||
//go:build unit
|
||||
|
||||
package service
|
||||
|
||||
import (
|
||||
"context"
|
||||
"os"
|
||||
"path/filepath"
|
||||
"testing"
|
||||
"time"
|
||||
|
||||
"github.com/Wei-Shaw/sub2api/internal/config"
|
||||
"github.com/Wei-Shaw/sub2api/internal/pkg/timezone"
|
||||
"github.com/stretchr/testify/require"
|
||||
)
|
||||
|
||||
// ---------------------------------------------------------------------------
|
||||
// deepseekPeakMultiplierAt：官方峰谷口径（2026-08-23 起生效）
|
||||
// 高峰时段 01:00–04:00 与 06:00–10:00 UTC（半开区间，仅工作日）；
|
||||
// 北京时间周六/周日全天低谷；高峰价 = 2× 低谷价。
|
||||
// 2026-08-24 为周一（工作日），2026-08-22 周六、2026-08-23 周日。
|
||||
// ---------------------------------------------------------------------------
|
||||
|
||||
func TestDeepseekPeakMultiplierAt(t *testing.T) {
|
||||
mon := func(hour, min int) time.Time { return time.Date(2026, 8, 24, hour, min, 0, 0, time.UTC) }
|
||||
sat := func(hour, min int) time.Time { return time.Date(2026, 8, 22, hour, min, 0, 0, time.UTC) }
|
||||
sun := func(hour, min int) time.Time { return time.Date(2026, 8, 23, hour, min, 0, 0, time.UTC) }
|
||||
|
||||
tests := []struct {
|
||||
name string
|
||||
now time.Time
|
||||
want float64
|
||||
}{
|
||||
// 工作日高峰窗口边界（半开区间）
|
||||
{"weekday 01:00 peak start", mon(1, 0), 2.0},
|
||||
{"weekday 03:59 peak upper bound", mon(3, 59), 2.0},
|
||||
{"weekday 04:00 peak end", mon(4, 0), 1.0},
|
||||
{"weekday 06:00 peak start", mon(6, 0), 2.0},
|
||||
{"weekday 09:59 peak upper bound", mon(9, 59), 2.0},
|
||||
{"weekday 10:00 peak end", mon(10, 0), 1.0},
|
||||
// 工作日低谷时段
|
||||
{"weekday 00:00 off-peak", mon(0, 0), 1.0},
|
||||
{"weekday 05:00 off-peak", mon(5, 0), 1.0},
|
||||
{"weekday 12:00 off-peak", mon(12, 0), 1.0},
|
||||
{"weekday 23:59 off-peak", mon(23, 59), 1.0},
|
||||
// 北京时间周末全天低谷（即使 UTC 处于高峰时段）
|
||||
{"saturday utc 02:00 beijing sat 10:00", sat(2, 0), 1.0},
|
||||
{"sunday utc 07:00 beijing sun 15:00", sun(7, 0), 1.0},
|
||||
// 北京时间与 UTC 跨日边界：UTC 周六 16:30 = 北京周日 00:30 → 周末低谷
|
||||
{"utc saturday 16:30 = beijing sunday 00:30", sat(16, 30), 1.0},
|
||||
}
|
||||
for _, tt := range tests {
|
||||
t.Run(tt.name, func(t *testing.T) {
|
||||
require.Equal(t, tt.want, deepseekPeakMultiplierAt(tt.now))
|
||||
})
|
||||
}
|
||||
}
|
||||
|
||||
func TestIsDeepSeekModel(t *testing.T) {
|
||||
deepseek := []string{
|
||||
"deepseek-v4-flash", "deepseek-v4-pro", "deepseek-v4-flash-vision-exp",
|
||||
"deepseek-chat", "deepseek-reasoner", "deepseek-v3-2-251201",
|
||||
"deepseek-coder", "deepseek-foo", "deepseek-v4-pro-0813",
|
||||
"DEEPSEEK-V4-PRO", " deepseek-v4-flash ",
|
||||
}
|
||||
for _, m := range deepseek {
|
||||
require.True(t, isDeepSeekModel(m), "model %q should be deepseek", m)
|
||||
}
|
||||
|
||||
nonDeepseek := []string{
|
||||
"gpt-5.4", "claude-sonnet-4", "deepseekcoder", // 无连字符不算 deepseek- 前缀
|
||||
"", " deepseek", // 无连字符后缀
|
||||
}
|
||||
for _, m := range nonDeepseek {
|
||||
require.False(t, isDeepSeekModel(m), "model %q should not be deepseek", m)
|
||||
}
|
||||
}
|
||||
|
||||
// ---------------------------------------------------------------------------
|
||||
// 默认价卡（Source=LiteLLM）按官方峰谷倍率计费；分组/渠道自定义定价不叠加
|
||||
// ---------------------------------------------------------------------------
|
||||
|
||||
func TestCalculateCostUnified_DeepseekDefaultCardPeakMultiplier(t *testing.T) {
|
||||
bs := newTestBillingService()
|
||||
resolver := NewModelPricingResolver(nil, bs)
|
||||
|
||||
tokens := UsageTokens{InputTokens: 1000, OutputTokens: 500, CacheReadTokens: 1000}
|
||||
// 低谷成本：1000*2.2e-7 + 500*6.6e-7 + 1000*7e-9 = 5.57e-4
|
||||
offPeakTotal := 1000*2.2e-7 + 500*6.6e-7 + 1000*7e-9
|
||||
|
||||
offPeak, err := bs.CalculateCostUnified(CostInput{
|
||||
Ctx: context.Background(), Model: "deepseek-v4-flash", Tokens: tokens,
|
||||
RateMultiplier: 1.0, Resolver: resolver,
|
||||
PricingAt: time.Date(2026, 8, 24, 12, 0, 0, 0, time.UTC), // 周一低谷
|
||||
})
|
||||
require.NoError(t, err)
|
||||
require.InDelta(t, offPeakTotal, offPeak.TotalCost, 1e-10)
|
||||
|
||||
peak, err := bs.CalculateCostUnified(CostInput{
|
||||
Ctx: context.Background(), Model: "deepseek-v4-flash", Tokens: tokens,
|
||||
RateMultiplier: 1.0, Resolver: resolver,
|
||||
PricingAt: time.Date(2026, 8, 24, 2, 0, 0, 0, time.UTC), // 周一高峰
|
||||
})
|
||||
require.NoError(t, err)
|
||||
require.InDelta(t, offPeakTotal*2, peak.TotalCost, 1e-10)
|
||||
}
|
||||
|
||||
func TestCalculateCostUnified_DeepseekProDefaultCardPeakMultiplier(t *testing.T) {
|
||||
bs := newTestBillingService()
|
||||
resolver := NewModelPricingResolver(nil, bs)
|
||||
|
||||
tokens := UsageTokens{InputTokens: 1000, OutputTokens: 500, CacheReadTokens: 1000}
|
||||
offPeakTotal := 1000*6.6e-7 + 500*1.98e-6 + 1000*2.2e-8
|
||||
|
||||
offPeak, err := bs.CalculateCostUnified(CostInput{
|
||||
Ctx: context.Background(), Model: "deepseek-v4-pro", Tokens: tokens,
|
||||
RateMultiplier: 1.0, Resolver: resolver,
|
||||
PricingAt: time.Date(2026, 8, 24, 12, 0, 0, 0, time.UTC),
|
||||
})
|
||||
require.NoError(t, err)
|
||||
require.InDelta(t, offPeakTotal, offPeak.TotalCost, 1e-10)
|
||||
|
||||
peak, err := bs.CalculateCostUnified(CostInput{
|
||||
Ctx: context.Background(), Model: "deepseek-v4-pro", Tokens: tokens,
|
||||
RateMultiplier: 1.0, Resolver: resolver,
|
||||
PricingAt: time.Date(2026, 8, 24, 6, 30, 0, 0, time.UTC), // 周一高峰
|
||||
})
|
||||
require.NoError(t, err)
|
||||
require.InDelta(t, offPeakTotal*2, peak.TotalCost, 1e-10)
|
||||
}
|
||||
|
||||
func TestCalculateCostUnified_DeepseekVersionedNamePeakMultiplier(t *testing.T) {
|
||||
bs := newTestBillingService()
|
||||
resolver := NewModelPricingResolver(nil, bs)
|
||||
|
||||
tokens := UsageTokens{InputTokens: 1000, OutputTokens: 500, CacheReadTokens: 1000}
|
||||
offPeakTotal := 1000*2.2e-7 + 500*6.6e-7 + 1000*7e-9
|
||||
|
||||
offPeak, err := bs.CalculateCostUnified(CostInput{
|
||||
Ctx: context.Background(), Model: "deepseek-v4-flash-0731", Tokens: tokens,
|
||||
RateMultiplier: 1.0, Resolver: resolver,
|
||||
PricingAt: time.Date(2026, 8, 24, 12, 0, 0, 0, time.UTC), // 周一低谷
|
||||
})
|
||||
require.NoError(t, err)
|
||||
require.InDelta(t, offPeakTotal, offPeak.TotalCost, 1e-10)
|
||||
|
||||
peak, err := bs.CalculateCostUnified(CostInput{
|
||||
Ctx: context.Background(), Model: "deepseek-v4-flash-0731", Tokens: tokens,
|
||||
RateMultiplier: 1.0, Resolver: resolver,
|
||||
PricingAt: time.Date(2026, 8, 24, 2, 0, 0, 0, time.UTC), // 周一高峰
|
||||
})
|
||||
require.NoError(t, err)
|
||||
require.InDelta(t, offPeakTotal*2, peak.TotalCost, 1e-10)
|
||||
}
|
||||
|
||||
func TestCalculateCostUnified_DeepseekGroupPricingNotScaledByPeak(t *testing.T) {
|
||||
bs := newTestBillingService()
|
||||
resolver := NewModelPricingResolver(nil, bs)
|
||||
|
||||
inputPrice := 1e-6
|
||||
outputPrice := 2e-6
|
||||
group := &Group{
|
||||
ID: 1, Name: "ds-group", Platform: PlatformDeepseek, Status: StatusActive,
|
||||
ModelPricing: []ChannelModelPricing{{
|
||||
Models: []string{"deepseek-v4-flash"}, BillingMode: BillingModeToken,
|
||||
InputPrice: &inputPrice, OutputPrice: &outputPrice,
|
||||
}},
|
||||
}
|
||||
resolved := resolver.Resolve(context.Background(), PricingInput{Model: "deepseek-v4-flash", Group: group})
|
||||
require.Equal(t, PricingSourceGroup, resolved.Source)
|
||||
|
||||
tokens := UsageTokens{InputTokens: 1000, OutputTokens: 500, CacheReadTokens: 1000}
|
||||
// 分组自定义价：1000*1e-6 + 500*2e-6 + 1000*7e-9（缓存读沿用官方 flash 价）
|
||||
groupTotal := 1000*1e-6 + 500*2e-6 + 1000*7e-9
|
||||
|
||||
for _, pricingAt := range []time.Time{
|
||||
time.Date(2026, 8, 24, 12, 0, 0, 0, time.UTC), // 低谷
|
||||
time.Date(2026, 8, 24, 2, 0, 0, 0, time.UTC), // 高峰
|
||||
} {
|
||||
cost, err := bs.CalculateCostUnified(CostInput{
|
||||
Ctx: context.Background(), Model: "deepseek-v4-flash", Group: group,
|
||||
Tokens: tokens, RateMultiplier: 1.0, Resolver: resolver, PricingAt: pricingAt,
|
||||
})
|
||||
require.NoError(t, err)
|
||||
require.InDelta(t, groupTotal, cost.TotalCost, 1e-10,
|
||||
"分组自定义定价不应叠加官方峰谷倍率（pricingAt=%v）", pricingAt)
|
||||
}
|
||||
}
|
||||
|
||||
func TestCalculateCostUnified_NonDeepseekDefaultCardNotScaledByPeak(t *testing.T) {
|
||||
bs := newTestBillingService()
|
||||
resolver := NewModelPricingResolver(nil, bs)
|
||||
|
||||
tokens := UsageTokens{InputTokens: 1000, OutputTokens: 500}
|
||||
total := 1000*3e-6 + 500*15e-6 // claude-sonnet-4 fallback
|
||||
|
||||
for _, pricingAt := range []time.Time{
|
||||
time.Date(2026, 8, 24, 12, 0, 0, 0, time.UTC),
|
||||
time.Date(2026, 8, 24, 2, 0, 0, 0, time.UTC),
|
||||
} {
|
||||
cost, err := bs.CalculateCostUnified(CostInput{
|
||||
Ctx: context.Background(), Model: "claude-sonnet-4", Tokens: tokens,
|
||||
RateMultiplier: 1.0, Resolver: resolver, PricingAt: pricingAt,
|
||||
})
|
||||
require.NoError(t, err)
|
||||
require.InDelta(t, total, cost.TotalCost, 1e-10,
|
||||
"非 DeepSeek 模型不应受官方峰谷倍率影响（pricingAt=%v）", pricingAt)
|
||||
}
|
||||
}
|
||||
|
||||
func TestCalculateCostUnified_DeepseekPricingAtZeroFallsBackToNow(t *testing.T) {
|
||||
bs := newTestBillingService()
|
||||
resolver := NewModelPricingResolver(nil, bs)
|
||||
|
||||
tokens := UsageTokens{InputTokens: 1000, OutputTokens: 500}
|
||||
base := CostInput{
|
||||
Ctx: context.Background(), Model: "deepseek-v4-flash", Tokens: tokens,
|
||||
RateMultiplier: 1.0, Resolver: resolver,
|
||||
}
|
||||
|
||||
// PricingAt 零值 → 回退 timezone.Now()，与显式传入当前时刻结果一致。
|
||||
costZero, err := bs.CalculateCostUnified(base)
|
||||
require.NoError(t, err)
|
||||
|
||||
costNow, err := bs.CalculateCostUnified(CostInput{
|
||||
Ctx: base.Ctx, Model: base.Model, Tokens: base.Tokens,
|
||||
RateMultiplier: base.RateMultiplier, Resolver: base.Resolver,
|
||||
PricingAt: timezone.Now(),
|
||||
})
|
||||
require.NoError(t, err)
|
||||
require.Equal(t, costZero.TotalCost, costNow.TotalCost)
|
||||
}
|
||||
|
||||
// ---------------------------------------------------------------------------
|
||||
// 官方价强制覆盖（远端旧价兜底）与未知 deepseek-* flash 兜底
|
||||
// ---------------------------------------------------------------------------
|
||||
|
||||
func TestGetModelPricing_DeepseekForcesOfficialRatesOverJSON(t *testing.T) {
|
||||
// JSON 给任意价（模拟远端旧价/占位价），deepseek-* 必须被强制覆盖为官方低谷价。
|
||||
pricingSvc := &PricingService{pricingData: map[string]*LiteLLMModelPricing{
|
||||
"deepseek-v4-flash": {InputCostPerToken: 1e-6, OutputCostPerToken: 2e-6, CacheReadInputTokenCost: 1e-8},
|
||||
"deepseek-v4-pro": {InputCostPerToken: 1e-6, OutputCostPerToken: 2e-6, CacheReadInputTokenCost: 1e-8},
|
||||
"deepseek-v4-flash-vision-exp": {InputCostPerToken: 1e-6, OutputCostPerToken: 2e-6, CacheReadInputTokenCost: 1e-8},
|
||||
"deepseek-chat": {InputCostPerToken: 1e-6, OutputCostPerToken: 2e-6, CacheReadInputTokenCost: 1e-8},
|
||||
"deepseek-reasoner": {InputCostPerToken: 1e-6, OutputCostPerToken: 2e-6, CacheReadInputTokenCost: 1e-8},
|
||||
}}
|
||||
bs := NewBillingService(&config.Config{}, pricingSvc)
|
||||
|
||||
tests := []struct {
|
||||
model string
|
||||
input, output, cacheRead float64
|
||||
}{
|
||||
{"deepseek-v4-flash", 2.2e-7, 6.6e-7, 7e-9},
|
||||
{"deepseek-v4-flash-vision-exp", 2.2e-7, 6.6e-7, 7e-9},
|
||||
{"deepseek-v4-pro", 6.6e-7, 1.98e-6, 2.2e-8},
|
||||
// 已停服的 chat/reasoner：即使 JSON 有旧条目也按 flash 价兜底。
|
||||
{"deepseek-chat", 2.2e-7, 6.6e-7, 7e-9},
|
||||
{"deepseek-reasoner", 2.2e-7, 6.6e-7, 7e-9},
|
||||
}
|
||||
for _, tt := range tests {
|
||||
t.Run(tt.model, func(t *testing.T) {
|
||||
pricing, err := bs.GetModelPricing(tt.model)
|
||||
require.NoError(t, err)
|
||||
require.InDelta(t, tt.input, pricing.InputPricePerToken, 1e-15)
|
||||
require.InDelta(t, tt.output, pricing.OutputPricePerToken, 1e-15)
|
||||
require.InDelta(t, tt.cacheRead, pricing.CacheReadPricePerToken, 1e-15)
|
||||
require.True(t, bs.HasIdentifiedTokenPricing(tt.model))
|
||||
})
|
||||
}
|
||||
|
||||
// 版本化名称（不在 JSON / fallbackPrices 精确表中）：按子串归档计价。
|
||||
versioned := []struct {
|
||||
model string
|
||||
input, output, cacheRead float64
|
||||
}{
|
||||
{"deepseek-v4-pro-0813", 6.6e-7, 1.98e-6, 2.2e-8},
|
||||
{"deepseek-v4-flash-0731", 2.2e-7, 6.6e-7, 7e-9},
|
||||
}
|
||||
for _, tt := range versioned {
|
||||
t.Run(tt.model, func(t *testing.T) {
|
||||
pricing, err := bs.GetModelPricing(tt.model)
|
||||
require.NoError(t, err)
|
||||
require.InDelta(t, tt.input, pricing.InputPricePerToken, 1e-15)
|
||||
require.InDelta(t, tt.output, pricing.OutputPricePerToken, 1e-15)
|
||||
require.InDelta(t, tt.cacheRead, pricing.CacheReadPricePerToken, 1e-15)
|
||||
})
|
||||
}
|
||||
}
|
||||
|
||||
func TestGetModelPricing_UnknownDeepseekMapsToFlash(t *testing.T) {
|
||||
// JSON 含 $0 占位条目（如旧 deepseek-v3-2-251201）：未知 deepseek-* 不再
|
||||
// fail-closed，统一按 flash 价兜底（2.2e-7/6.6e-7/7e-9），不得按 $0 计费。
|
||||
pricingSvc := &PricingService{pricingData: map[string]*LiteLLMModelPricing{
|
||||
"deepseek-v3-2-251201": {InputCostPerToken: 0, OutputCostPerToken: 0},
|
||||
}}
|
||||
bs := NewBillingService(&config.Config{}, pricingSvc)
|
||||
|
||||
for _, m := range []string{"deepseek-v3-2-251201", "deepseek-chat", "deepseek-reasoner", "deepseek-foo"} {
|
||||
t.Run(m, func(t *testing.T) {
|
||||
pricing, err := bs.GetModelPricing(m)
|
||||
require.NoError(t, err)
|
||||
require.InDelta(t, 2.2e-7, pricing.InputPricePerToken, 1e-15)
|
||||
require.InDelta(t, 6.6e-7, pricing.OutputPricePerToken, 1e-15)
|
||||
require.InDelta(t, 7e-9, pricing.CacheReadPricePerToken, 1e-15)
|
||||
})
|
||||
}
|
||||
}
|
||||
|
||||
// ---------------------------------------------------------------------------
|
||||
// 本地兜底 JSON：无 $0 占位条目，官方模型价格为官方低谷价
|
||||
// ---------------------------------------------------------------------------
|
||||
|
||||
func TestDeepseekPricingFileMatchesOfficialRates(t *testing.T) {
|
||||
data, err := os.ReadFile(filepath.Join("..", "..", "resources", "model-pricing", "model_prices_and_context_window.json"))
|
||||
require.NoError(t, err)
|
||||
|
||||
pricingSvc := &PricingService{}
|
||||
pricingData, err := pricingSvc.parsePricingData(data)
|
||||
require.NoError(t, err)
|
||||
|
||||
_, ok := pricingData["deepseek-v3-2-251201"]
|
||||
require.False(t, ok, "deepseek-v3-2-251201（$0 占位条目）必须从价格表中移除")
|
||||
for _, discontinued := range []string{"deepseek-chat", "deepseek-reasoner"} {
|
||||
_, ok := pricingData[discontinued]
|
||||
require.False(t, ok, "%s 已停止服务，必须从价格表中移除", discontinued)
|
||||
}
|
||||
|
||||
tests := []struct {
|
||||
model string
|
||||
input, output, cacheRead float64
|
||||
}{
|
||||
{"deepseek-v4-flash", 2.2e-7, 6.6e-7, 7e-9},
|
||||
{"deepseek-v4-flash-vision-exp", 2.2e-7, 6.6e-7, 7e-9},
|
||||
{"deepseek-v4-pro", 6.6e-7, 1.98e-6, 2.2e-8},
|
||||
}
|
||||
for _, tt := range tests {
|
||||
t.Run(tt.model, func(t *testing.T) {
|
||||
entry, ok := pricingData[tt.model]
|
||||
require.True(t, ok, "model %s must exist in pricing file", tt.model)
|
||||
require.InDelta(t, tt.input, entry.InputCostPerToken, 1e-15)
|
||||
require.InDelta(t, tt.output, entry.OutputCostPerToken, 1e-15)
|
||||
require.InDelta(t, tt.cacheRead, entry.CacheReadInputTokenCost, 1e-15)
|
||||
})
|
||||
}
|
||||
}
|
||||
@@ -745,15 +745,15 @@
|
||||
"supports_web_search": true,
|
||||
"supports_xhigh_reasoning_effort": true
|
||||
},
|
||||
"deepseek-chat": {
|
||||
"cache_read_input_token_cost": 2.8e-08,
|
||||
"input_cost_per_token": 2.8e-07,
|
||||
"deepseek-v4-flash": {
|
||||
"cache_read_input_token_cost": 7e-09,
|
||||
"input_cost_per_token": 2.2e-07,
|
||||
"litellm_provider": "deepseek",
|
||||
"max_input_tokens": 131072,
|
||||
"max_output_tokens": 8192,
|
||||
"max_tokens": 8192,
|
||||
"max_input_tokens": 1048576,
|
||||
"max_output_tokens": 393216,
|
||||
"max_tokens": 393216,
|
||||
"mode": "chat",
|
||||
"output_cost_per_token": 4.2e-07,
|
||||
"output_cost_per_token": 6.6e-07,
|
||||
"source": "https://api-docs.deepseek.com/quick_start/pricing",
|
||||
"supported_endpoints": [
|
||||
"/v1/chat/completions"
|
||||
@@ -762,44 +762,54 @@
|
||||
"supports_native_streaming": true,
|
||||
"supports_parallel_function_calling": true,
|
||||
"supports_prompt_caching": true,
|
||||
"supports_reasoning": true,
|
||||
"supports_response_schema": true,
|
||||
"supports_system_messages": true,
|
||||
"supports_tool_choice": true
|
||||
},
|
||||
"deepseek-reasoner": {
|
||||
"cache_read_input_token_cost": 2.8e-08,
|
||||
"input_cost_per_token": 2.8e-07,
|
||||
"deepseek-v4-flash-vision-exp": {
|
||||
"cache_read_input_token_cost": 7e-09,
|
||||
"input_cost_per_token": 2.2e-07,
|
||||
"litellm_provider": "deepseek",
|
||||
"max_input_tokens": 131072,
|
||||
"max_output_tokens": 65536,
|
||||
"max_tokens": 65536,
|
||||
"max_input_tokens": 1048576,
|
||||
"max_output_tokens": 393216,
|
||||
"max_tokens": 393216,
|
||||
"mode": "chat",
|
||||
"output_cost_per_token": 4.2e-07,
|
||||
"output_cost_per_token": 6.6e-07,
|
||||
"source": "https://api-docs.deepseek.com/quick_start/pricing",
|
||||
"supported_endpoints": [
|
||||
"/v1/chat/completions"
|
||||
],
|
||||
"supports_function_calling": false,
|
||||
"supports_function_calling": true,
|
||||
"supports_native_streaming": true,
|
||||
"supports_parallel_function_calling": false,
|
||||
"supports_parallel_function_calling": true,
|
||||
"supports_prompt_caching": true,
|
||||
"supports_reasoning": true,
|
||||
"supports_response_schema": true,
|
||||
"supports_system_messages": true,
|
||||
"supports_tool_choice": false
|
||||
"supports_tool_choice": true,
|
||||
"supports_vision": true
|
||||
},
|
||||
"deepseek-v3-2-251201": {
|
||||
"input_cost_per_token": 0.0,
|
||||
"litellm_provider": "volcengine",
|
||||
"max_input_tokens": 98304,
|
||||
"max_output_tokens": 32768,
|
||||
"max_tokens": 32768,
|
||||
"deepseek-v4-pro": {
|
||||
"cache_read_input_token_cost": 2.2e-08,
|
||||
"input_cost_per_token": 6.6e-07,
|
||||
"litellm_provider": "deepseek",
|
||||
"max_input_tokens": 1048576,
|
||||
"max_output_tokens": 393216,
|
||||
"max_tokens": 393216,
|
||||
"mode": "chat",
|
||||
"output_cost_per_token": 0.0,
|
||||
"supports_assistant_prefill": true,
|
||||
"output_cost_per_token": 1.98e-06,
|
||||
"source": "https://api-docs.deepseek.com/quick_start/pricing",
|
||||
"supported_endpoints": [
|
||||
"/v1/chat/completions"
|
||||
],
|
||||
"supports_function_calling": true,
|
||||
"supports_native_streaming": true,
|
||||
"supports_parallel_function_calling": true,
|
||||
"supports_prompt_caching": true,
|
||||
"supports_reasoning": true,
|
||||
"supports_response_schema": true,
|
||||
"supports_system_messages": true,
|
||||
"supports_tool_choice": true
|
||||
},
|
||||
"gemini-2.0-flash": {
|
||||
|
||||
@@ -109,7 +109,8 @@ const qwenModels = [
|
||||
|
||||
// DeepSeek
|
||||
const deepseekModels = [
|
||||
'deepseek-chat', 'deepseek-coder', 'deepseek-reasoner',
|
||||
'deepseek-v4-pro', 'deepseek-v4-flash', 'deepseek-v4-flash-vision-exp',
|
||||
'deepseek-coder',
|
||||
'deepseek-v3', 'deepseek-v3-0324',
|
||||
'deepseek-r1', 'deepseek-r1-0528',
|
||||
'deepseek-r1-distill-qwen-32b', 'deepseek-r1-distill-qwen-14b', 'deepseek-r1-distill-qwen-7b',
|
||||
|
||||
Reference in New Issue
Block a user