Merge remote-tracking branch 'origin/dev-next' into dev-next

# Conflicts:
#	go.sum

docs/changelog.md

@@ -2,9 +2,9 @@
 icon: material/alert-decagram
 ---
 
-#### 1.13.0-rc.6
+#### 1.13.0-rc.7
 
-* Fixes and improvements
+* Add advertise tags support for Tailscale endpoint
 
 Important changes since 1.12:
 
@@ -136,6 +136,7 @@ See [Dial Fields](/configuration/shared/dial/#bind_address_no_port).
 
 Tailscale endpoint can now create a system TUN interface to handle traffic directly.
 New `relay_server_port` and `relay_server_static_endpoints` options for incoming relay connections.
+New `advertise_tags` option for ACL tag advertisement.
 
 See [Tailscale endpoint](/configuration/endpoint/tailscale/).
 

docs/configuration/endpoint/tailscale.md

@@ -9,6 +9,7 @@ icon: material/new-box
     :material-plus: [system_interface](#system_interface)  
     :material-plus: [system_interface_name](#system_interface_name)  
     :material-plus: [system_interface_mtu](#system_interface_mtu)  
+    :material-plus: [advertise_tags](#advertise_tags)
 
 !!! question "Since sing-box 1.12.0"
 
@@ -28,6 +29,7 @@ icon: material/new-box
   "exit_node_allow_lan_access": false,
   "advertise_routes": [],
   "advertise_exit_node": false,
+  "advertise_tags": [],
   "relay_server_port": 0,
   "relay_server_static_endpoints": [],
   "system_interface": false,
@@ -102,6 +104,14 @@ Example: `["192.168.1.1/24"]`
 
 Indicates whether the node should advertise itself as an exit node.
 
+#### advertise_tags
+
+!!! question "Since sing-box 1.13.0"
+
+Tags to advertise for this node, for ACL enforcement purposes.
+
+Example: `["tag:server"]`
+
 #### relay_server_port
 
 !!! question "Since sing-box 1.13.0"

docs/configuration/endpoint/tailscale.zh.md

@@ -9,6 +9,7 @@ icon: material/new-box
     :material-plus: [system_interface](#system_interface)  
     :material-plus: [system_interface_name](#system_interface_name)  
     :material-plus: [system_interface_mtu](#system_interface_mtu)  
+    :material-plus: [advertise_tags](#advertise_tags)
 
 !!! question "自 sing-box 1.12.0 起"
 
@@ -28,6 +29,7 @@ icon: material/new-box
   "exit_node_allow_lan_access": false,
   "advertise_routes": [],
   "advertise_exit_node": false,
+  "advertise_tags": [],
   "relay_server_port": 0,
   "relay_server_static_endpoints": [],
   "system_interface": false,
@@ -101,6 +103,14 @@ icon: material/new-box
 
 指示节点是否应将自己通告为出口节点。
 
+#### advertise_tags
+
+!!! question "自 sing-box 1.13.0 起"
+
+为此节点通告的标签，用于 ACL 执行。
+
+示例：`["tag:server"]`
+
 #### relay_server_port
 
 !!! question "自 sing-box 1.13.0 起"

go.mod

@@ -3,7 +3,7 @@ module github.com/sagernet/sing-box
 go 1.24.7
 
 require (
-	github.com/anthropics/anthropic-sdk-go v1.19.0
+	github.com/anthropics/anthropic-sdk-go v1.26.0
 	github.com/anytls/sing-anytls v0.0.11
 	github.com/caddyserver/certmagic v0.25.0
 	github.com/coder/websocket v1.8.14
@@ -23,7 +23,7 @@ require (
 	github.com/metacubex/utls v1.8.4
 	github.com/mholt/acmez/v3 v3.1.4
 	github.com/miekg/dns v1.1.69
-	github.com/openai/openai-go/v3 v3.15.0
+	github.com/openai/openai-go/v3 v3.23.0
 	github.com/oschwald/maxminddb-golang v1.13.1
 	github.com/sagernet/asc-go v0.0.0-20241217030726-d563060fe4e1
 	github.com/sagernet/bbolt v0.0.0-20231014093535-ea5cb2fe9f0a
@@ -44,7 +44,7 @@ require (
 	github.com/sagernet/sing-vmess v0.2.8-0.20250909125414-3aed155119a1
 	github.com/sagernet/smux v1.5.50-sing-box-mod.1
 	github.com/sagernet/tailscale v1.92.4-sing-box-1.13-mod.6
-	github.com/sagernet/wireguard-go v0.0.2-beta.1.0.20250917110311-16510ac47288
+	github.com/sagernet/wireguard-go v0.0.2-beta.1.0.20260224074747-506b7631853c
 	github.com/sagernet/ws v0.0.0-20231204124109-acfe8907c854
 	github.com/spf13/cobra v1.10.2
 	github.com/stretchr/testify v1.11.1

go.sum

@@ -8,8 +8,8 @@ github.com/alexbrainman/sspi v0.0.0-20231016080023-1a75b4708caa h1:LHTHcTQiSGT7V
 github.com/alexbrainman/sspi v0.0.0-20231016080023-1a75b4708caa/go.mod h1:cEWa1LVoE5KvSD9ONXsZrj0z6KqySlCCNKHlLzbqAt4=
 github.com/andybalholm/brotli v1.1.0 h1:eLKJA0d02Lf0mVpIDgYnqXcUn0GqVmEFny3VuID1U3M=
 github.com/andybalholm/brotli v1.1.0/go.mod h1:sms7XGricyQI9K10gOSf56VKKWS4oLer58Q+mhRPtnY=
-github.com/anthropics/anthropic-sdk-go v1.19.0 h1:mO6E+ffSzLRvR/YUH9KJC0uGw0uV8GjISIuzem//3KE=
+github.com/anthropics/anthropic-sdk-go v1.26.0 h1:oUTzFaUpAevfuELAP1sjL6CQJ9HHAfT7CoSYSac11PY=
-github.com/anthropics/anthropic-sdk-go v1.19.0/go.mod h1:WTz31rIUHUHqai2UslPpw5CwXrQP3geYBioRV4WOLvE=
+github.com/anthropics/anthropic-sdk-go v1.26.0/go.mod h1:qUKmaW+uuPB64iy1l+4kOSvaLqPXnHTTBKH6RVZ7q5Q=
 github.com/anytls/sing-anytls v0.0.11 h1:w8e9Uj1oP3m4zxkyZDewPk0EcQbvVxb7Nn+rapEx4fc=
 github.com/anytls/sing-anytls v0.0.11/go.mod h1:7rjN6IukwysmdusYsrV51Fgu1uW6vsrdd6ctjnEAln8=
 github.com/benbjohnson/clock v1.1.0/go.mod h1:J11/hYXuz8f4ySSvYwY0FKfm+ezbsZBKZxNJlLklBHA=
@@ -40,6 +40,8 @@ github.com/dblohm7/wingoes v0.0.0-20240119213807-a09d6be7affa h1:h8TfIT1xc8FWbww
 github.com/dblohm7/wingoes v0.0.0-20240119213807-a09d6be7affa/go.mod h1:Nx87SkVqTKd8UtT+xu7sM/l+LgXs6c0aHrlKusR+2EQ=
 github.com/dgrijalva/jwt-go/v4 v4.0.0-preview1 h1:CaO/zOnF8VvUfEbhRatPcwKVWamvbYd8tQGRWacE9kU=
 github.com/dgrijalva/jwt-go/v4 v4.0.0-preview1/go.mod h1:+hnT3ywWDTAFrW5aE+u2Sa/wT555ZqwoCS+pk3p6ry4=
+github.com/dnaeon/go-vcr v1.2.0 h1:zHCHvJYTMh1N7xnV7zf1m1GPBF9Ad0Jk/whtQ1663qI=
+github.com/dnaeon/go-vcr v1.2.0/go.mod h1:R4UdLID7HZT3taECzJs4YgbbH6PIGXB6W/sc5OLb6RQ=
 github.com/ebitengine/purego v0.9.1 h1:a/k2f2HQU3Pi399RPW1MOaZyhKJL9w/xFpKAg4q1s0A=
 github.com/ebitengine/purego v0.9.1/go.mod h1:iIjxzd6CiRiOG0UyXP+V1+jWqUXVjPKLAI0mRfJZTmQ=
 github.com/florianl/go-nfqueue/v2 v2.0.2 h1:FL5lQTeetgpCvac1TRwSfgaXUn0YSO7WzGvWNIp3JPE=
@@ -141,8 +143,8 @@ github.com/mitchellh/go-ps v1.0.0 h1:i6ampVEEF4wQFF+bkYfwYgY+F/uYJDktmvLPf7qIgjc
 github.com/mitchellh/go-ps v1.0.0/go.mod h1:J4lOc8z8yJs6vUwklHw2XEIiT4z4C40KtWVN3nvg8Pg=
 github.com/nfnt/resize v0.0.0-20180221191011-83c6a9932646 h1:zYyBkD/k9seD2A7fsi6Oo2LfFZAehjjQMERAvZLEDnQ=
 github.com/nfnt/resize v0.0.0-20180221191011-83c6a9932646/go.mod h1:jpp1/29i3P1S/RLdc7JQKbRpFeM1dOBd8T9ki5s+AY8=
-github.com/openai/openai-go/v3 v3.15.0 h1:hk99rM7YPz+M99/5B/zOQcVwFRLLMdprVGx1vaZ8XMo=
+github.com/openai/openai-go/v3 v3.23.0 h1:FRFwTcB4FoWFtIunTY/8fgHvzSHgqbfWjiCwOMVrsvw=
-github.com/openai/openai-go/v3 v3.15.0/go.mod h1:cdufnVK14cWcT9qA1rRtrXx4FTRsgbDPW7Ia7SS5cZo=
+github.com/openai/openai-go/v3 v3.23.0/go.mod h1:cdufnVK14cWcT9qA1rRtrXx4FTRsgbDPW7Ia7SS5cZo=
 github.com/oschwald/maxminddb-golang v1.13.1 h1:G3wwjdN9JmIK2o/ermkHM+98oX5fS+k5MbwsmL4MRQE=
 github.com/oschwald/maxminddb-golang v1.13.1/go.mod h1:K4pgV9N/GcK694KSTmVSDTODk4IsCNThNdTmnaBZ/F8=
 github.com/pierrec/lz4/v4 v4.1.21 h1:yOVMLb6qSIDP67pl/5F7RepeKYu/VmTyEXvuMI5d9mQ=
@@ -261,8 +263,6 @@ github.com/sagernet/sing-shadowsocks2 v0.2.1 h1:dWV9OXCeFPuYGHb6IRqlSptVnSzOelnq
 github.com/sagernet/sing-shadowsocks2 v0.2.1/go.mod h1:RnXS0lExcDAovvDeniJ4IKa2IuChrdipolPYWBv9hWQ=
 github.com/sagernet/sing-shadowtls v0.2.1-0.20250503051639-fcd445d33c11 h1:tK+75l64tm9WvEFrYRE1t0YxoFdWQqw/h7Uhzj0vJ+w=
 github.com/sagernet/sing-shadowtls v0.2.1-0.20250503051639-fcd445d33c11/go.mod h1:sWqKnGlMipCHaGsw1sTTlimyUpgzP4WP3pjhCsYt9oA=
-github.com/sagernet/sing-tun v0.8.0-beta.17.0.20260223095246-5715a3919a7f h1:3B8auqVameLvPhWxzw5S8QWdLTv19o0M8LYxPOXNm0g=
-github.com/sagernet/sing-tun v0.8.0-beta.17.0.20260223095246-5715a3919a7f/go.mod h1:+HAK/y9GZljdT0KYKMYDR8MjjqnqDDQZYp5ZZQoRzS8=
 github.com/sagernet/sing-tun v0.8.0-beta.18 h1:C6oHxP9BNBVEVdC9ABMTXmKej9mUVtcuw2v+IiBS8yw=
 github.com/sagernet/sing-tun v0.8.0-beta.18/go.mod h1:+HAK/y9GZljdT0KYKMYDR8MjjqnqDDQZYp5ZZQoRzS8=
 github.com/sagernet/sing-vmess v0.2.8-0.20250909125414-3aed155119a1 h1:aSwUNYUkVyVvdmBSufR8/nRFonwJeKSIROxHcm5br9o=
@@ -271,8 +271,8 @@ github.com/sagernet/smux v1.5.50-sing-box-mod.1 h1:XkJcivBC9V4wBjiGXIXZ229aZCU1h
 github.com/sagernet/smux v1.5.50-sing-box-mod.1/go.mod h1:NjhsCEWedJm7eFLyhuBgIEzwfhRmytrUoiLluxs5Sk8=
 github.com/sagernet/tailscale v1.92.4-sing-box-1.13-mod.6 h1:eYz/OpMqWCvO2++iw3dEuzrlfC2xv78GdlGvprIM6O8=
 github.com/sagernet/tailscale v1.92.4-sing-box-1.13-mod.6/go.mod h1:m87GAn4UcesHQF3leaPFEINZETO5za1LGn1GJdNDgNc=
-github.com/sagernet/wireguard-go v0.0.2-beta.1.0.20250917110311-16510ac47288 h1:E2tZFeg9mGYGQ7E7BbxMv1cU35HxwgRm6tPKI2Pp7DA=
+github.com/sagernet/wireguard-go v0.0.2-beta.1.0.20260224074747-506b7631853c h1:f9cXNB+IOOPnR8DOLMTpr42jf7naxh5Un5Y09BBf5Cg=
-github.com/sagernet/wireguard-go v0.0.2-beta.1.0.20250917110311-16510ac47288/go.mod h1:WUxgxUDZoCF2sxVmW+STSxatP02Qn3FcafTiI2BLtE0=
+github.com/sagernet/wireguard-go v0.0.2-beta.1.0.20260224074747-506b7631853c/go.mod h1:WUxgxUDZoCF2sxVmW+STSxatP02Qn3FcafTiI2BLtE0=
 github.com/sagernet/ws v0.0.0-20231204124109-acfe8907c854 h1:6uUiZcDRnZSAegryaUGwPC/Fj13JSHwiTftrXhMmYOc=
 github.com/sagernet/ws v0.0.0-20231204124109-acfe8907c854/go.mod h1:LtfoSK3+NG57tvnVEHgcuBW9ujgE8enPSgzgwStwCAA=
 github.com/shopspring/decimal v1.2.0 h1:abSATXmQEYyShuxI4/vyW3tV1MrKAJzCZ/0zLUXYbsQ=
@@ -434,6 +434,8 @@ gopkg.in/check.v1 v1.0.0-20180628173108-788fd7840127/go.mod h1:Co6ibVJAznAaIkqp8
 gopkg.in/natefinch/lumberjack.v2 v2.2.1 h1:bBRl1b0OH9s/DuPhuXpNl+VtCaJXFZ5/uEFST95x9zc=
 gopkg.in/natefinch/lumberjack.v2 v2.2.1/go.mod h1:YD8tP3GAjkrDg1eZH7EGmyESg/lsYskCTPBJVb9jqSc=
 gopkg.in/yaml.v1 v1.0.0-20140924161607-9f9df34309c0/go.mod h1:WDnlLJ4WF5VGsH/HVa3CI79GS0ol3YnhVnKP89i0kNg=
+gopkg.in/yaml.v2 v2.2.8 h1:obN1ZagJSUGI0Ek/LBmuj4SNLPfIny3KsKFopxRdj10=
+gopkg.in/yaml.v2 v2.2.8/go.mod h1:hI93XBmqTisBFMUTm0b8Fm+jr3Dg1NNxqwp+5A1VGuI=
 gopkg.in/yaml.v2 v2.2.2/go.mod h1:hI93XBmqTisBFMUTm0b8Fm+jr3Dg1NNxqwp+5A1VGuI=
 gopkg.in/yaml.v2 v2.2.8/go.mod h1:hI93XBmqTisBFMUTm0b8Fm+jr3Dg1NNxqwp+5A1VGuI=
 gopkg.in/yaml.v3 v3.0.0-20200313102051-9f266ea9e77c/go.mod h1:K4uyk7z7BCEPqu6E+C64Yfv1cQ7kz7rIZviUmN+EgEM=

option/tailscale.go

@@ -22,6 +22,7 @@ type TailscaleEndpointOptions struct {
 	ExitNodeAllowLANAccess     bool                       `json:"exit_node_allow_lan_access,omitempty"`
 	AdvertiseRoutes            []netip.Prefix             `json:"advertise_routes,omitempty"`
 	AdvertiseExitNode          bool                       `json:"advertise_exit_node,omitempty"`
+	AdvertiseTags              badoption.Listable[string] `json:"advertise_tags,omitempty"`
 	RelayServerPort            *uint16                    `json:"relay_server_port,omitempty"`
 	RelayServerStaticEndpoints []netip.AddrPort           `json:"relay_server_static_endpoints,omitempty"`
 	SystemInterface            bool                       `json:"system_interface,omitempty"`

protocol/tailscale/endpoint.go

@@ -97,6 +97,7 @@ type Endpoint struct {
 	exitNodeAllowLANAccess     bool
 	advertiseRoutes            []netip.Prefix
 	advertiseExitNode          bool
+	advertiseTags              []string
 	relayServerPort            *uint16
 	relayServerStaticEndpoints []netip.AddrPort
 
@@ -244,6 +245,7 @@ func NewEndpoint(ctx context.Context, router adapter.Router, logger log.ContextL
 		exitNodeAllowLANAccess:     options.ExitNodeAllowLANAccess,
 		advertiseRoutes:            options.AdvertiseRoutes,
 		advertiseExitNode:          options.AdvertiseExitNode,
+		advertiseTags:              options.AdvertiseTags,
 		relayServerPort:            options.RelayServerPort,
 		relayServerStaticEndpoints: options.RelayServerStaticEndpoints,
 		udpTimeout:                 udpTimeout,
@@ -360,24 +362,24 @@ func (t *Endpoint) Start(stage adapter.StartStage) error {
 	perfs := &ipn.MaskedPrefs{
 		Prefs: ipn.Prefs{
 			RouteAll:        t.acceptRoutes,
+			AdvertiseRoutes: t.advertiseRoutes,
+			AdvertiseTags:   t.advertiseTags,
 		},
 		RouteAllSet:                   true,
 		ExitNodeIPSet:                 true,
 		AdvertiseRoutesSet:            true,
-	}
+		AdvertiseTagsSet:              true,
-	if len(t.advertiseRoutes) > 0 {
+		RelayServerPortSet:            true,
-		perfs.AdvertiseRoutes = t.advertiseRoutes
+		RelayServerStaticEndpointsSet: true,
 	}
 	if t.advertiseExitNode {
 		perfs.AdvertiseRoutes = append(perfs.AdvertiseRoutes, tsaddr.ExitRoutes()...)
 	}
 	if t.relayServerPort != nil {
 		perfs.RelayServerPort = t.relayServerPort
-		perfs.RelayServerPortSet = true
 	}
 	if len(t.relayServerStaticEndpoints) > 0 {
 		perfs.RelayServerStaticEndpoints = t.relayServerStaticEndpoints
-		perfs.RelayServerStaticEndpointsSet = true
 	}
 	_, err = localBackend.EditPrefs(perfs)
 	if err != nil {

service/ccm/service.go

@@ -425,6 +425,8 @@ func (s *Service) handleResponseWithTracking(writer http.ResponseWriter, respons
 					usage.OutputTokens,
 					usage.CacheReadInputTokens,
 					usage.CacheCreationInputTokens,
+					usage.CacheCreation.Ephemeral5mInputTokens,
+					usage.CacheCreation.Ephemeral1hInputTokens,
 					username,
 				)
 			}
@@ -485,6 +487,8 @@ func (s *Service) handleResponseWithTracking(writer http.ResponseWriter, respons
 							accumulatedUsage.InputTokens = messageStart.Message.Usage.InputTokens
 							accumulatedUsage.CacheReadInputTokens = messageStart.Message.Usage.CacheReadInputTokens
 							accumulatedUsage.CacheCreationInputTokens = messageStart.Message.Usage.CacheCreationInputTokens
+							accumulatedUsage.CacheCreation.Ephemeral5mInputTokens = messageStart.Message.Usage.CacheCreation.Ephemeral5mInputTokens
+							accumulatedUsage.CacheCreation.Ephemeral1hInputTokens = messageStart.Message.Usage.CacheCreation.Ephemeral1hInputTokens
 						}
 					case "message_delta":
 						messageDelta := event.AsMessageDelta()
@@ -519,6 +523,8 @@ func (s *Service) handleResponseWithTracking(writer http.ResponseWriter, respons
 						accumulatedUsage.OutputTokens,
 						accumulatedUsage.CacheReadInputTokens,
 						accumulatedUsage.CacheCreationInputTokens,
+						accumulatedUsage.CacheCreation.Ephemeral5mInputTokens,
+						accumulatedUsage.CacheCreation.Ephemeral1hInputTokens,
 						username,
 					)
 				}

service/ccm/service_usage.go

@@ -19,6 +19,8 @@ type UsageStats struct {
 	OutputTokens                    int64 `json:"output_tokens"`
 	CacheReadInputTokens            int64 `json:"cache_read_input_tokens"`
 	CacheCreationInputTokens        int64 `json:"cache_creation_input_tokens"`
+	CacheCreation5MinuteInputTokens int64 `json:"cache_creation_5m_input_tokens,omitempty"`
+	CacheCreation1HourInputTokens   int64 `json:"cache_creation_1h_input_tokens,omitempty"`
 }
 
 type CostCombination struct {
@@ -47,6 +49,8 @@ type UsageStatsJSON struct {
 	OutputTokens                    int64   `json:"output_tokens"`
 	CacheReadInputTokens            int64   `json:"cache_read_input_tokens"`
 	CacheCreationInputTokens        int64   `json:"cache_creation_input_tokens"`
+	CacheCreation5MinuteInputTokens int64   `json:"cache_creation_5m_input_tokens,omitempty"`
+	CacheCreation1HourInputTokens   int64   `json:"cache_creation_1h_input_tokens,omitempty"`
 	CostUSD                         float64 `json:"cost_usd"`
 }
 
@@ -72,7 +76,8 @@ type ModelPricing struct {
 	InputPrice             float64
 	OutputPrice            float64
 	CacheReadPrice         float64
-	CacheWritePrice float64
+	CacheWritePrice5Minute float64
+	CacheWritePrice1Hour   float64
 }
 
 type modelFamily struct {
@@ -82,143 +87,205 @@ type modelFamily struct {
 }
 
 var (
-	opus4Pricing = ModelPricing{
+	opus46StandardPricing = ModelPricing{
-		InputPrice:      15.0,
+		InputPrice:             5.0,
-		OutputPrice:     75.0,
+		OutputPrice:            25.0,
-		CacheReadPrice:  1.5,
+		CacheReadPrice:         0.5,
-		CacheWritePrice: 18.75,
+		CacheWritePrice5Minute: 6.25,
+		CacheWritePrice1Hour:   10.0,
 	}
 
-	sonnet4StandardPricing = ModelPricing{
+	opus46PremiumPricing = ModelPricing{
-		InputPrice:      3.0,
+		InputPrice:             10.0,
-		OutputPrice:     15.0,
+		OutputPrice:            37.5,
-		CacheReadPrice:  0.3,
+		CacheReadPrice:         1.0,
-		CacheWritePrice: 3.75,
+		CacheWritePrice5Minute: 12.5,
-	}
+		CacheWritePrice1Hour:   20.0,
-
-	sonnet4PremiumPricing = ModelPricing{
-		InputPrice:      6.0,
-		OutputPrice:     22.5,
-		CacheReadPrice:  0.6,
-		CacheWritePrice: 7.5,
-	}
-
-	haiku4Pricing = ModelPricing{
-		InputPrice:      1.0,
-		OutputPrice:     5.0,
-		CacheReadPrice:  0.1,
-		CacheWritePrice: 1.25,
-	}
-
-	haiku35Pricing = ModelPricing{
-		InputPrice:      0.8,
-		OutputPrice:     4.0,
-		CacheReadPrice:  0.08,
-		CacheWritePrice: 1.0,
-	}
-
-	sonnet35Pricing = ModelPricing{
-		InputPrice:      3.0,
-		OutputPrice:     15.0,
-		CacheReadPrice:  0.3,
-		CacheWritePrice: 3.75,
 	}
 
 	opus45Pricing = ModelPricing{
 		InputPrice:             5.0,
 		OutputPrice:            25.0,
 		CacheReadPrice:         0.5,
-		CacheWritePrice: 6.25,
+		CacheWritePrice5Minute: 6.25,
+		CacheWritePrice1Hour:   10.0,
+	}
+
+	opus4Pricing = ModelPricing{
+		InputPrice:             15.0,
+		OutputPrice:            75.0,
+		CacheReadPrice:         1.5,
+		CacheWritePrice5Minute: 18.75,
+		CacheWritePrice1Hour:   30.0,
+	}
+
+	sonnet46StandardPricing = ModelPricing{
+		InputPrice:             3.0,
+		OutputPrice:            15.0,
+		CacheReadPrice:         0.3,
+		CacheWritePrice5Minute: 3.75,
+		CacheWritePrice1Hour:   6.0,
+	}
+
+	sonnet46PremiumPricing = ModelPricing{
+		InputPrice:             6.0,
+		OutputPrice:            22.5,
+		CacheReadPrice:         0.6,
+		CacheWritePrice5Minute: 7.5,
+		CacheWritePrice1Hour:   12.0,
 	}
 
 	sonnet45StandardPricing = ModelPricing{
 		InputPrice:             3.0,
 		OutputPrice:            15.0,
 		CacheReadPrice:         0.3,
-		CacheWritePrice: 3.75,
+		CacheWritePrice5Minute: 3.75,
+		CacheWritePrice1Hour:   6.0,
 	}
 
 	sonnet45PremiumPricing = ModelPricing{
 		InputPrice:             6.0,
 		OutputPrice:            22.5,
 		CacheReadPrice:         0.6,
-		CacheWritePrice: 7.5,
+		CacheWritePrice5Minute: 7.5,
+		CacheWritePrice1Hour:   12.0,
+	}
+
+	sonnet4StandardPricing = ModelPricing{
+		InputPrice:             3.0,
+		OutputPrice:            15.0,
+		CacheReadPrice:         0.3,
+		CacheWritePrice5Minute: 3.75,
+		CacheWritePrice1Hour:   6.0,
+	}
+
+	sonnet4PremiumPricing = ModelPricing{
+		InputPrice:             6.0,
+		OutputPrice:            22.5,
+		CacheReadPrice:         0.6,
+		CacheWritePrice5Minute: 7.5,
+		CacheWritePrice1Hour:   12.0,
+	}
+
+	sonnet37Pricing = ModelPricing{
+		InputPrice:             3.0,
+		OutputPrice:            15.0,
+		CacheReadPrice:         0.3,
+		CacheWritePrice5Minute: 3.75,
+		CacheWritePrice1Hour:   6.0,
+	}
+
+	sonnet35Pricing = ModelPricing{
+		InputPrice:             3.0,
+		OutputPrice:            15.0,
+		CacheReadPrice:         0.3,
+		CacheWritePrice5Minute: 3.75,
+		CacheWritePrice1Hour:   6.0,
 	}
 
 	haiku45Pricing = ModelPricing{
 		InputPrice:             1.0,
 		OutputPrice:            5.0,
 		CacheReadPrice:         0.1,
-		CacheWritePrice: 1.25,
+		CacheWritePrice5Minute: 1.25,
+		CacheWritePrice1Hour:   2.0,
+	}
+
+	haiku4Pricing = ModelPricing{
+		InputPrice:             1.0,
+		OutputPrice:            5.0,
+		CacheReadPrice:         0.1,
+		CacheWritePrice5Minute: 1.25,
+		CacheWritePrice1Hour:   2.0,
+	}
+
+	haiku35Pricing = ModelPricing{
+		InputPrice:             0.8,
+		OutputPrice:            4.0,
+		CacheReadPrice:         0.08,
+		CacheWritePrice5Minute: 1.0,
+		CacheWritePrice1Hour:   1.6,
 	}
 
 	haiku3Pricing = ModelPricing{
 		InputPrice:             0.25,
 		OutputPrice:            1.25,
 		CacheReadPrice:         0.03,
-		CacheWritePrice: 0.3,
+		CacheWritePrice5Minute: 0.3,
+		CacheWritePrice1Hour:   0.5,
 	}
 
 	opus3Pricing = ModelPricing{
 		InputPrice:             15.0,
 		OutputPrice:            75.0,
 		CacheReadPrice:         1.5,
-		CacheWritePrice: 18.75,
+		CacheWritePrice5Minute: 18.75,
+		CacheWritePrice1Hour:   30.0,
 	}
 
 	modelFamilies = []modelFamily{
 		{
-			pattern:         regexp.MustCompile(`^claude-opus-4-5-`),
+			pattern:         regexp.MustCompile(`^claude-opus-4-6(?:-|$)`),
+			standardPricing: opus46StandardPricing,
+			premiumPricing:  &opus46PremiumPricing,
+		},
+		{
+			pattern:         regexp.MustCompile(`^claude-opus-4-5(?:-|$)`),
 			standardPricing: opus45Pricing,
 			premiumPricing:  nil,
 		},
 		{
-			pattern:         regexp.MustCompile(`^claude-(?:opus-4-|4-opus-|opus-4-1-)`),
+			pattern:         regexp.MustCompile(`^claude-(?:opus-4(?:-|$)|4-opus-)`),
 			standardPricing: opus4Pricing,
 			premiumPricing:  nil,
 		},
 		{
-			pattern:         regexp.MustCompile(`^claude-(?:opus-3-|3-opus-)`),
+			pattern:         regexp.MustCompile(`^claude-(?:opus-3(?:-|$)|3-opus-)`),
 			standardPricing: opus3Pricing,
 			premiumPricing:  nil,
 		},
 		{
-			pattern:         regexp.MustCompile(`^claude-(?:sonnet-4-5-|4-5-sonnet-)`),
+			pattern:         regexp.MustCompile(`^claude-(?:sonnet-4-6(?:-|$)|4-6-sonnet-)`),
+			standardPricing: sonnet46StandardPricing,
+			premiumPricing:  &sonnet46PremiumPricing,
+		},
+		{
+			pattern:         regexp.MustCompile(`^claude-(?:sonnet-4-5(?:-|$)|4-5-sonnet-)`),
 			standardPricing: sonnet45StandardPricing,
 			premiumPricing:  &sonnet45PremiumPricing,
 		},
 		{
-			pattern:         regexp.MustCompile(`^claude-3-7-sonnet-`),
+			pattern:         regexp.MustCompile(`^claude-(?:sonnet-4(?:-|$)|4-sonnet-)`),
 			standardPricing: sonnet4StandardPricing,
 			premiumPricing:  &sonnet4PremiumPricing,
 		},
 		{
-			pattern:         regexp.MustCompile(`^claude-(?:sonnet-4-|4-sonnet-)`),
+			pattern:         regexp.MustCompile(`^claude-3-7-sonnet(?:-|$)`),
-			standardPricing: sonnet4StandardPricing,
+			standardPricing: sonnet37Pricing,
-			premiumPricing:  &sonnet4PremiumPricing,
+			premiumPricing:  nil,
 		},
 		{
-			pattern:         regexp.MustCompile(`^claude-3-5-sonnet-`),
+			pattern:         regexp.MustCompile(`^claude-3-5-sonnet(?:-|$)`),
 			standardPricing: sonnet35Pricing,
 			premiumPricing:  nil,
 		},
 		{
-			pattern:         regexp.MustCompile(`^claude-(?:haiku-4-5-|4-5-haiku-)`),
+			pattern:         regexp.MustCompile(`^claude-(?:haiku-4-5(?:-|$)|4-5-haiku-)`),
 			standardPricing: haiku45Pricing,
 			premiumPricing:  nil,
 		},
 		{
-			pattern:         regexp.MustCompile(`^claude-haiku-4-`),
+			pattern:         regexp.MustCompile(`^claude-haiku-4(?:-|$)`),
 			standardPricing: haiku4Pricing,
 			premiumPricing:  nil,
 		},
 		{
-			pattern:         regexp.MustCompile(`^claude-3-5-haiku-`),
+			pattern:         regexp.MustCompile(`^claude-3-5-haiku(?:-|$)`),
 			standardPricing: haiku35Pricing,
 			premiumPricing:  nil,
 		},
 		{
-			pattern:         regexp.MustCompile(`^claude-3-haiku-`),
+			pattern:         regexp.MustCompile(`^claude-3-haiku(?:-|$)`),
 			standardPricing: haiku3Pricing,
 			premiumPricing:  nil,
 		},
@@ -243,10 +310,20 @@ func getPricing(model string, contextWindow int) ModelPricing {
 func calculateCost(stats UsageStats, model string, contextWindow int) float64 {
 	pricing := getPricing(model, contextWindow)
 
+	cacheCreationCost := 0.0
+	if stats.CacheCreation5MinuteInputTokens > 0 || stats.CacheCreation1HourInputTokens > 0 {
+		cacheCreationCost =
+			float64(stats.CacheCreation5MinuteInputTokens)*pricing.CacheWritePrice5Minute +
+				float64(stats.CacheCreation1HourInputTokens)*pricing.CacheWritePrice1Hour
+	} else {
+		// Backward compatibility for usage files generated before TTL split tracking.
+		cacheCreationCost = float64(stats.CacheCreationInputTokens) * pricing.CacheWritePrice5Minute
+	}
+
 	cost := (float64(stats.InputTokens)*pricing.InputPrice +
 		float64(stats.OutputTokens)*pricing.OutputPrice +
 		float64(stats.CacheReadInputTokens)*pricing.CacheReadPrice +
-		float64(stats.CacheCreationInputTokens)*pricing.CacheWritePrice) / 1_000_000
+		cacheCreationCost) / 1_000_000
 
 	return math.Round(cost*100) / 100
 }
@@ -279,6 +356,8 @@ func (u *AggregatedUsage) ToJSON() *AggregatedUsageJSON {
 				OutputTokens:                    combo.Total.OutputTokens,
 				CacheReadInputTokens:            combo.Total.CacheReadInputTokens,
 				CacheCreationInputTokens:        combo.Total.CacheCreationInputTokens,
+				CacheCreation5MinuteInputTokens: combo.Total.CacheCreation5MinuteInputTokens,
+				CacheCreation1HourInputTokens:   combo.Total.CacheCreation1HourInputTokens,
 				CostUSD:                         totalCost,
 			},
 			ByUser: make(map[string]UsageStatsJSON),
@@ -295,6 +374,8 @@ func (u *AggregatedUsage) ToJSON() *AggregatedUsageJSON {
 				OutputTokens:                    userStats.OutputTokens,
 				CacheReadInputTokens:            userStats.CacheReadInputTokens,
 				CacheCreationInputTokens:        userStats.CacheCreationInputTokens,
+				CacheCreation5MinuteInputTokens: userStats.CacheCreation5MinuteInputTokens,
+				CacheCreation1HourInputTokens:   userStats.CacheCreation1HourInputTokens,
 				CostUSD:                         userCost,
 			}
 		}
@@ -367,7 +448,13 @@ func (u *AggregatedUsage) Save() error {
 	return err
 }
 
-func (u *AggregatedUsage) AddUsage(model string, contextWindow int, messagesCount int, inputTokens, outputTokens, cacheReadTokens, cacheCreationTokens int64, user string) error {
+func (u *AggregatedUsage) AddUsage(
+	model string,
+	contextWindow int,
+	messagesCount int,
+	inputTokens, outputTokens, cacheReadTokens, cacheCreationTokens, cacheCreation5MinuteTokens, cacheCreation1HourTokens int64,
+	user string,
+) error {
 	if model == "" {
 		return E.New("model cannot be empty")
 	}
@@ -400,6 +487,10 @@ func (u *AggregatedUsage) AddUsage(model string, contextWindow int, messagesCoun
 		combo = &u.Combinations[len(u.Combinations)-1]
 	}
 
+	if cacheCreationTokens == 0 {
+		cacheCreationTokens = cacheCreation5MinuteTokens + cacheCreation1HourTokens
+	}
+
 	// Update total stats
 	combo.Total.RequestCount++
 	combo.Total.MessagesCount += messagesCount
@@ -407,6 +498,8 @@ func (u *AggregatedUsage) AddUsage(model string, contextWindow int, messagesCoun
 	combo.Total.OutputTokens += outputTokens
 	combo.Total.CacheReadInputTokens += cacheReadTokens
 	combo.Total.CacheCreationInputTokens += cacheCreationTokens
+	combo.Total.CacheCreation5MinuteInputTokens += cacheCreation5MinuteTokens
+	combo.Total.CacheCreation1HourInputTokens += cacheCreation1HourTokens
 
 	// Update per-user stats if user is specified
 	if user != "" {
@@ -417,6 +510,8 @@ func (u *AggregatedUsage) AddUsage(model string, contextWindow int, messagesCoun
 		userStats.OutputTokens += outputTokens
 		userStats.CacheReadInputTokens += cacheReadTokens
 		userStats.CacheCreationInputTokens += cacheCreationTokens
+		userStats.CacheCreation5MinuteInputTokens += cacheCreation5MinuteTokens
+		userStats.CacheCreation1HourInputTokens += cacheCreation1HourTokens
 		combo.ByUser[user] = userStats
 	}
 

service/ocm/service.go

@@ -406,7 +406,9 @@ func (s *Service) handleResponseWithTracking(writer http.ResponseWriter, respons
 	isChatCompletions := path == "/v1/chat/completions"
 	mediaType, _, err := mime.ParseMediaType(response.Header.Get("Content-Type"))
 	isStreaming := err == nil && mediaType == "text/event-stream"
-
+	if !isStreaming && !isChatCompletions && response.Header.Get("Content-Type") == "" {
+		isStreaming = true
+	}
 	if !isStreaming {
 		bodyBytes, err := io.ReadAll(response.Body)
 		if err != nil {
@@ -414,13 +416,14 @@ func (s *Service) handleResponseWithTracking(writer http.ResponseWriter, respons
 			return
 		}
 
-		var responseModel string
+		var responseModel, serviceTier string
 		var inputTokens, outputTokens, cachedTokens int64
 
 		if isChatCompletions {
 			var chatCompletion openai.ChatCompletion
 			if json.Unmarshal(bodyBytes, &chatCompletion) == nil {
 				responseModel = chatCompletion.Model
+				serviceTier = string(chatCompletion.ServiceTier)
 				inputTokens = chatCompletion.Usage.PromptTokens
 				outputTokens = chatCompletion.Usage.CompletionTokens
 				cachedTokens = chatCompletion.Usage.PromptTokensDetails.CachedTokens
@@ -429,6 +432,7 @@ func (s *Service) handleResponseWithTracking(writer http.ResponseWriter, respons
 			var responsesResponse responses.Response
 			if json.Unmarshal(bodyBytes, &responsesResponse) == nil {
 				responseModel = string(responsesResponse.Model)
+				serviceTier = string(responsesResponse.ServiceTier)
 				inputTokens = responsesResponse.Usage.InputTokens
 				outputTokens = responsesResponse.Usage.OutputTokens
 				cachedTokens = responsesResponse.Usage.InputTokensDetails.CachedTokens
@@ -440,7 +444,7 @@ func (s *Service) handleResponseWithTracking(writer http.ResponseWriter, respons
 				responseModel = requestModel
 			}
 			if responseModel != "" {
-				s.usageTracker.AddUsage(responseModel, inputTokens, outputTokens, cachedTokens, username)
+				s.usageTracker.AddUsage(responseModel, inputTokens, outputTokens, cachedTokens, serviceTier, username)
 			}
 		}
 
@@ -455,7 +459,7 @@ func (s *Service) handleResponseWithTracking(writer http.ResponseWriter, respons
 	}
 
 	var inputTokens, outputTokens, cachedTokens int64
-	var responseModel string
+	var responseModel, serviceTier string
 	buffer := make([]byte, buf.BufferSize)
 	var leftover []byte
 
@@ -490,6 +494,9 @@ func (s *Service) handleResponseWithTracking(writer http.ResponseWriter, respons
 							if chatChunk.Model != "" {
 								responseModel = chatChunk.Model
 							}
+							if chatChunk.ServiceTier != "" {
+								serviceTier = string(chatChunk.ServiceTier)
+							}
 							if chatChunk.Usage.PromptTokens > 0 {
 								inputTokens = chatChunk.Usage.PromptTokens
 								cachedTokens = chatChunk.Usage.PromptTokensDetails.CachedTokens
@@ -506,6 +513,9 @@ func (s *Service) handleResponseWithTracking(writer http.ResponseWriter, respons
 								if string(completedEvent.Response.Model) != "" {
 									responseModel = string(completedEvent.Response.Model)
 								}
+								if completedEvent.Response.ServiceTier != "" {
+									serviceTier = string(completedEvent.Response.ServiceTier)
+								}
 								if completedEvent.Response.Usage.InputTokens > 0 {
 									inputTokens = completedEvent.Response.Usage.InputTokens
 									cachedTokens = completedEvent.Response.Usage.InputTokensDetails.CachedTokens
@@ -534,7 +544,7 @@ func (s *Service) handleResponseWithTracking(writer http.ResponseWriter, respons
 
 			if inputTokens > 0 || outputTokens > 0 {
 				if responseModel != "" {
-					s.usageTracker.AddUsage(responseModel, inputTokens, outputTokens, cachedTokens, username)
+					s.usageTracker.AddUsage(responseModel, inputTokens, outputTokens, cachedTokens, serviceTier, username)
 				}
 			}
 			return

service/ocm/service_usage.go

@@ -5,6 +5,7 @@ import (
 	"math"
 	"os"
 	"regexp"
+	"strings"
 	"sync"
 	"time"
 
@@ -43,6 +44,7 @@ func (u *UsageStats) UnmarshalJSON(data []byte) error {
 
 type CostCombination struct {
 	Model       string                `json:"model"`
+	ServiceTier string                `json:"service_tier,omitempty"`
 	Total       UsageStats            `json:"total"`
 	ByUser      map[string]UsageStats `json:"by_user"`
 }
@@ -69,6 +71,7 @@ type UsageStatsJSON struct {
 
 type CostCombinationJSON struct {
 	Model       string                    `json:"model"`
+	ServiceTier string                    `json:"service_tier,omitempty"`
 	Total       UsageStatsJSON            `json:"total"`
 	ByUser      map[string]UsageStatsJSON `json:"by_user"`
 }
@@ -95,7 +98,123 @@ type modelFamily struct {
 	pricing ModelPricing
 }
 
+const (
+	serviceTierAuto     = "auto"
+	serviceTierDefault  = "default"
+	serviceTierFlex     = "flex"
+	serviceTierPriority = "priority"
+	serviceTierScale    = "scale"
+)
+
 var (
+	gpt52Pricing = ModelPricing{
+		InputPrice:       1.75,
+		OutputPrice:      14.0,
+		CachedInputPrice: 0.175,
+	}
+
+	gpt5Pricing = ModelPricing{
+		InputPrice:       1.25,
+		OutputPrice:      10.0,
+		CachedInputPrice: 0.125,
+	}
+
+	gpt5MiniPricing = ModelPricing{
+		InputPrice:       0.25,
+		OutputPrice:      2.0,
+		CachedInputPrice: 0.025,
+	}
+
+	gpt5NanoPricing = ModelPricing{
+		InputPrice:       0.05,
+		OutputPrice:      0.4,
+		CachedInputPrice: 0.005,
+	}
+
+	gpt52CodexPricing = ModelPricing{
+		InputPrice:       1.75,
+		OutputPrice:      14.0,
+		CachedInputPrice: 0.175,
+	}
+
+	gpt51CodexPricing = ModelPricing{
+		InputPrice:       1.25,
+		OutputPrice:      10.0,
+		CachedInputPrice: 0.125,
+	}
+
+	gpt51CodexMiniPricing = ModelPricing{
+		InputPrice:       0.25,
+		OutputPrice:      2.0,
+		CachedInputPrice: 0.025,
+	}
+
+	gpt52ProPricing = ModelPricing{
+		InputPrice:       21.0,
+		OutputPrice:      168.0,
+		CachedInputPrice: 21.0,
+	}
+
+	gpt5ProPricing = ModelPricing{
+		InputPrice:       15.0,
+		OutputPrice:      120.0,
+		CachedInputPrice: 15.0,
+	}
+
+	gpt52FlexPricing = ModelPricing{
+		InputPrice:       0.875,
+		OutputPrice:      7.0,
+		CachedInputPrice: 0.0875,
+	}
+
+	gpt5FlexPricing = ModelPricing{
+		InputPrice:       0.625,
+		OutputPrice:      5.0,
+		CachedInputPrice: 0.0625,
+	}
+
+	gpt5MiniFlexPricing = ModelPricing{
+		InputPrice:       0.125,
+		OutputPrice:      1.0,
+		CachedInputPrice: 0.0125,
+	}
+
+	gpt5NanoFlexPricing = ModelPricing{
+		InputPrice:       0.025,
+		OutputPrice:      0.2,
+		CachedInputPrice: 0.0025,
+	}
+
+	gpt52PriorityPricing = ModelPricing{
+		InputPrice:       3.5,
+		OutputPrice:      28.0,
+		CachedInputPrice: 0.35,
+	}
+
+	gpt5PriorityPricing = ModelPricing{
+		InputPrice:       2.5,
+		OutputPrice:      20.0,
+		CachedInputPrice: 0.25,
+	}
+
+	gpt5MiniPriorityPricing = ModelPricing{
+		InputPrice:       0.45,
+		OutputPrice:      3.6,
+		CachedInputPrice: 0.045,
+	}
+
+	gpt52CodexPriorityPricing = ModelPricing{
+		InputPrice:       3.5,
+		OutputPrice:      28.0,
+		CachedInputPrice: 0.35,
+	}
+
+	gpt51CodexPriorityPricing = ModelPricing{
+		InputPrice:       2.5,
+		OutputPrice:      20.0,
+		CachedInputPrice: 0.25,
+	}
+
 	gpt4oPricing = ModelPricing{
 		InputPrice:       2.5,
 		OutputPrice:      10.0,
@@ -111,7 +230,19 @@ var (
 	gpt4oAudioPricing = ModelPricing{
 		InputPrice:       2.5,
 		OutputPrice:      10.0,
-		CachedInputPrice: 1.25,
+		CachedInputPrice: 2.5,
+	}
+
+	gpt4oMiniAudioPricing = ModelPricing{
+		InputPrice:       0.15,
+		OutputPrice:      0.6,
+		CachedInputPrice: 0.15,
+	}
+
+	gptAudioMiniPricing = ModelPricing{
+		InputPrice:       0.6,
+		OutputPrice:      2.4,
+		CachedInputPrice: 0.6,
 	}
 
 	o1Pricing = ModelPricing{
@@ -120,6 +251,12 @@ var (
 		CachedInputPrice: 7.5,
 	}
 
+	o1ProPricing = ModelPricing{
+		InputPrice:       150.0,
+		OutputPrice:      600.0,
+		CachedInputPrice: 150.0,
+	}
+
 	o1MiniPricing = ModelPricing{
 		InputPrice:       1.1,
 		OutputPrice:      4.4,
@@ -135,13 +272,55 @@ var (
 	o3Pricing = ModelPricing{
 		InputPrice:       2.0,
 		OutputPrice:      8.0,
-		CachedInputPrice: 1.0,
+		CachedInputPrice: 0.5,
+	}
+
+	o3ProPricing = ModelPricing{
+		InputPrice:       20.0,
+		OutputPrice:      80.0,
+		CachedInputPrice: 20.0,
+	}
+
+	o3DeepResearchPricing = ModelPricing{
+		InputPrice:       10.0,
+		OutputPrice:      40.0,
+		CachedInputPrice: 2.5,
 	}
 
 	o4MiniPricing = ModelPricing{
 		InputPrice:       1.1,
 		OutputPrice:      4.4,
-		CachedInputPrice: 0.55,
+		CachedInputPrice: 0.275,
+	}
+
+	o4MiniDeepResearchPricing = ModelPricing{
+		InputPrice:       2.0,
+		OutputPrice:      8.0,
+		CachedInputPrice: 0.5,
+	}
+
+	o3FlexPricing = ModelPricing{
+		InputPrice:       1.0,
+		OutputPrice:      4.0,
+		CachedInputPrice: 0.25,
+	}
+
+	o4MiniFlexPricing = ModelPricing{
+		InputPrice:       0.55,
+		OutputPrice:      2.2,
+		CachedInputPrice: 0.138,
+	}
+
+	o3PriorityPricing = ModelPricing{
+		InputPrice:       3.5,
+		OutputPrice:      14.0,
+		CachedInputPrice: 0.875,
+	}
+
+	o4MiniPriorityPricing = ModelPricing{
+		InputPrice:       2.0,
+		OutputPrice:      8.0,
+		CachedInputPrice: 0.5,
 	}
 
 	gpt41Pricing = ModelPricing{
@@ -162,69 +341,358 @@ var (
 		CachedInputPrice: 0.025,
 	}
 
-	modelFamilies = []modelFamily{
+	gpt41PriorityPricing = ModelPricing{
+		InputPrice:       3.5,
+		OutputPrice:      14.0,
+		CachedInputPrice: 0.875,
+	}
+
+	gpt41MiniPriorityPricing = ModelPricing{
+		InputPrice:       0.7,
+		OutputPrice:      2.8,
+		CachedInputPrice: 0.175,
+	}
+
+	gpt41NanoPriorityPricing = ModelPricing{
+		InputPrice:       0.2,
+		OutputPrice:      0.8,
+		CachedInputPrice: 0.05,
+	}
+
+	gpt4oPriorityPricing = ModelPricing{
+		InputPrice:       4.25,
+		OutputPrice:      17.0,
+		CachedInputPrice: 2.125,
+	}
+
+	gpt4oMiniPriorityPricing = ModelPricing{
+		InputPrice:       0.25,
+		OutputPrice:      1.0,
+		CachedInputPrice: 0.125,
+	}
+
+	standardModelFamilies = []modelFamily{
 		{
-			pattern: regexp.MustCompile(`^gpt-4\.1-nano`),
+			pattern: regexp.MustCompile(`^gpt-5\.2-codex(?:$|-)`),
-			pricing: gpt41NanoPricing,
+			pricing: gpt52CodexPricing,
 		},
 		{
-			pattern: regexp.MustCompile(`^gpt-4\.1-mini`),
+			pattern: regexp.MustCompile(`^gpt-5\.1-codex-max(?:$|-)`),
-			pricing: gpt41MiniPricing,
+			pricing: gpt51CodexPricing,
 		},
 		{
-			pattern: regexp.MustCompile(`^gpt-4\.1`),
+			pattern: regexp.MustCompile(`^gpt-5\.1-codex-mini(?:$|-)`),
-			pricing: gpt41Pricing,
+			pricing: gpt51CodexMiniPricing,
 		},
 		{
-			pattern: regexp.MustCompile(`^o4-mini`),
+			pattern: regexp.MustCompile(`^gpt-5\.1-codex(?:$|-)`),
+			pricing: gpt51CodexPricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^gpt-5-codex(?:$|-)`),
+			pricing: gpt51CodexPricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^gpt-5\.2-chat-latest$`),
+			pricing: gpt52Pricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^gpt-5\.1-chat-latest$`),
+			pricing: gpt5Pricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^gpt-5-chat-latest$`),
+			pricing: gpt5Pricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^gpt-5\.2-pro(?:$|-)`),
+			pricing: gpt52ProPricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^gpt-5-pro(?:$|-)`),
+			pricing: gpt5ProPricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^gpt-5-mini(?:$|-)`),
+			pricing: gpt5MiniPricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^gpt-5-nano(?:$|-)`),
+			pricing: gpt5NanoPricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^gpt-5\.2(?:$|-)`),
+			pricing: gpt52Pricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^gpt-5\.1(?:$|-)`),
+			pricing: gpt5Pricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^gpt-5(?:$|-)`),
+			pricing: gpt5Pricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^o4-mini-deep-research(?:$|-)`),
+			pricing: o4MiniDeepResearchPricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^o4-mini(?:$|-)`),
 			pricing: o4MiniPricing,
 		},
 		{
-			pattern: regexp.MustCompile(`^o3-mini`),
+			pattern: regexp.MustCompile(`^o3-pro(?:$|-)`),
+			pricing: o3ProPricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^o3-deep-research(?:$|-)`),
+			pricing: o3DeepResearchPricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^o3-mini(?:$|-)`),
 			pricing: o3MiniPricing,
 		},
 		{
-			pattern: regexp.MustCompile(`^o3`),
+			pattern: regexp.MustCompile(`^o3(?:$|-)`),
 			pricing: o3Pricing,
 		},
 		{
-			pattern: regexp.MustCompile(`^o1-mini`),
+			pattern: regexp.MustCompile(`^o1-pro(?:$|-)`),
+			pricing: o1ProPricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^o1-mini(?:$|-)`),
 			pricing: o1MiniPricing,
 		},
 		{
-			pattern: regexp.MustCompile(`^o1`),
+			pattern: regexp.MustCompile(`^o1(?:$|-)`),
 			pricing: o1Pricing,
 		},
 		{
-			pattern: regexp.MustCompile(`^gpt-4o-audio`),
+			pattern: regexp.MustCompile(`^gpt-4o-mini-audio(?:$|-)`),
+			pricing: gpt4oMiniAudioPricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^gpt-audio-mini(?:$|-)`),
+			pricing: gptAudioMiniPricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^(?:gpt-4o-audio|gpt-audio)(?:$|-)`),
 			pricing: gpt4oAudioPricing,
 		},
 		{
-			pattern: regexp.MustCompile(`^gpt-4o-mini`),
+			pattern: regexp.MustCompile(`^gpt-4\.1-nano(?:$|-)`),
+			pricing: gpt41NanoPricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^gpt-4\.1-mini(?:$|-)`),
+			pricing: gpt41MiniPricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^gpt-4\.1(?:$|-)`),
+			pricing: gpt41Pricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^gpt-4o-mini(?:$|-)`),
 			pricing: gpt4oMiniPricing,
 		},
 		{
-			pattern: regexp.MustCompile(`^gpt-4o`),
+			pattern: regexp.MustCompile(`^gpt-4o(?:$|-)`),
 			pricing: gpt4oPricing,
 		},
 		{
-			pattern: regexp.MustCompile(`^chatgpt-4o`),
+			pattern: regexp.MustCompile(`^chatgpt-4o(?:$|-)`),
 			pricing: gpt4oPricing,
 		},
 	}
+
+	flexModelFamilies = []modelFamily{
+		{
+			pattern: regexp.MustCompile(`^gpt-5-mini(?:$|-)`),
+			pricing: gpt5MiniFlexPricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^gpt-5-nano(?:$|-)`),
+			pricing: gpt5NanoFlexPricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^gpt-5\.2(?:$|-)`),
+			pricing: gpt52FlexPricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^gpt-5\.1(?:$|-)`),
+			pricing: gpt5FlexPricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^gpt-5(?:$|-)`),
+			pricing: gpt5FlexPricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^o4-mini(?:$|-)`),
+			pricing: o4MiniFlexPricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^o3(?:$|-)`),
+			pricing: o3FlexPricing,
+		},
+	}
+
+	priorityModelFamilies = []modelFamily{
+		{
+			pattern: regexp.MustCompile(`^gpt-5\.2-codex(?:$|-)`),
+			pricing: gpt52CodexPriorityPricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^gpt-5\.1-codex-max(?:$|-)`),
+			pricing: gpt51CodexPriorityPricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^gpt-5\.1-codex(?:$|-)`),
+			pricing: gpt51CodexPriorityPricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^gpt-5-codex(?:$|-)`),
+			pricing: gpt51CodexPriorityPricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^gpt-5-mini(?:$|-)`),
+			pricing: gpt5MiniPriorityPricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^gpt-5\.2(?:$|-)`),
+			pricing: gpt52PriorityPricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^gpt-5\.1(?:$|-)`),
+			pricing: gpt5PriorityPricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^gpt-5(?:$|-)`),
+			pricing: gpt5PriorityPricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^o4-mini(?:$|-)`),
+			pricing: o4MiniPriorityPricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^o3(?:$|-)`),
+			pricing: o3PriorityPricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^gpt-4\.1-nano(?:$|-)`),
+			pricing: gpt41NanoPriorityPricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^gpt-4\.1-mini(?:$|-)`),
+			pricing: gpt41MiniPriorityPricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^gpt-4\.1(?:$|-)`),
+			pricing: gpt41PriorityPricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^gpt-4o-mini(?:$|-)`),
+			pricing: gpt4oMiniPriorityPricing,
+		},
+		{
+			pattern: regexp.MustCompile(`^gpt-4o(?:$|-)`),
+			pricing: gpt4oPriorityPricing,
+		},
+	}
 )
 
-func getPricing(model string) ModelPricing {
+func modelFamiliesForTier(serviceTier string) []modelFamily {
+	switch serviceTier {
+	case serviceTierFlex:
+		return flexModelFamilies
+	case serviceTierPriority:
+		return priorityModelFamilies
+	default:
+		return standardModelFamilies
+	}
+}
+
+func findPricingInFamilies(model string, modelFamilies []modelFamily) (ModelPricing, bool) {
 	for _, family := range modelFamilies {
 		if family.pattern.MatchString(model) {
-			return family.pricing
+			return family.pricing, true
 		}
 	}
+	return ModelPricing{}, false
+}
+
+func normalizeServiceTier(serviceTier string) string {
+	switch strings.ToLower(strings.TrimSpace(serviceTier)) {
+	case "", serviceTierAuto, serviceTierDefault:
+		return serviceTierDefault
+	case serviceTierFlex:
+		return serviceTierFlex
+	case serviceTierPriority:
+		return serviceTierPriority
+	case serviceTierScale:
+		// Scale-tier requests are prepaid differently and not listed in this usage file.
+		return serviceTierDefault
+	default:
+		return serviceTierDefault
+	}
+}
+
+func getPricing(model string, serviceTier string) ModelPricing {
+	normalizedServiceTier := normalizeServiceTier(serviceTier)
+	modelFamilies := modelFamiliesForTier(normalizedServiceTier)
+
+	if pricing, found := findPricingInFamilies(model, modelFamilies); found {
+		return pricing
+	}
+
+	normalizedModel := normalizeGPT5Model(model)
+	if normalizedModel != model {
+		if pricing, found := findPricingInFamilies(normalizedModel, modelFamilies); found {
+			return pricing
+		}
+	}
+
+	if normalizedServiceTier != serviceTierDefault {
+		if pricing, found := findPricingInFamilies(model, standardModelFamilies); found {
+			return pricing
+		}
+		if normalizedModel != model {
+			if pricing, found := findPricingInFamilies(normalizedModel, standardModelFamilies); found {
+				return pricing
+			}
+		}
+	}
+
 	return gpt4oPricing
 }
 
-func calculateCost(stats UsageStats, model string) float64 {
+func normalizeGPT5Model(model string) string {
-	pricing := getPricing(model)
+	if !strings.HasPrefix(model, "gpt-5.") {
+		return model
+	}
+
+	switch {
+	case strings.Contains(model, "-codex-mini"):
+		return "gpt-5.1-codex-mini"
+	case strings.Contains(model, "-codex-max"):
+		return "gpt-5.1-codex-max"
+	case strings.Contains(model, "-codex"):
+		return "gpt-5.2-codex"
+	case strings.Contains(model, "-chat-latest"):
+		return "gpt-5.2-chat-latest"
+	case strings.Contains(model, "-pro"):
+		return "gpt-5.2-pro"
+	case strings.Contains(model, "-mini"):
+		return "gpt-5-mini"
+	case strings.Contains(model, "-nano"):
+		return "gpt-5-nano"
+	default:
+		return "gpt-5.2"
+	}
+}
+
+func calculateCost(stats UsageStats, model string, serviceTier string) float64 {
+	pricing := getPricing(model, serviceTier)
 
 	regularInputTokens := stats.InputTokens - stats.CachedTokens
 	if regularInputTokens < 0 {
@@ -252,12 +720,13 @@ func (u *AggregatedUsage) ToJSON() *AggregatedUsageJSON {
 	}
 
 	for i, combo := range u.Combinations {
-		totalCost := calculateCost(combo.Total, combo.Model)
+		totalCost := calculateCost(combo.Total, combo.Model, combo.ServiceTier)
 
 		result.Costs.TotalUSD += totalCost
 
 		comboJSON := CostCombinationJSON{
 			Model:       combo.Model,
+			ServiceTier: combo.ServiceTier,
 			Total: UsageStatsJSON{
 				RequestCount: combo.Total.RequestCount,
 				InputTokens:  combo.Total.InputTokens,
@@ -269,7 +738,7 @@ func (u *AggregatedUsage) ToJSON() *AggregatedUsageJSON {
 		}
 
 		for user, userStats := range combo.ByUser {
-			userCost := calculateCost(userStats, combo.Model)
+			userCost := calculateCost(userStats, combo.Model, combo.ServiceTier)
 			result.Costs.ByUser[user] += userCost
 
 			comboJSON.ByUser[user] = UsageStatsJSON{
@@ -318,6 +787,7 @@ func (u *AggregatedUsage) Load() error {
 	u.Combinations = temp.Combinations
 
 	for i := range u.Combinations {
+		u.Combinations[i].ServiceTier = normalizeServiceTier(u.Combinations[i].ServiceTier)
 		if u.Combinations[i].ByUser == nil {
 			u.Combinations[i].ByUser = make(map[string]UsageStats)
 		}
@@ -349,11 +819,13 @@ func (u *AggregatedUsage) Save() error {
 	return err
 }
 
-func (u *AggregatedUsage) AddUsage(model string, inputTokens, outputTokens, cachedTokens int64, user string) error {
+func (u *AggregatedUsage) AddUsage(model string, inputTokens, outputTokens, cachedTokens int64, serviceTier string, user string) error {
 	if model == "" {
 		return E.New("model cannot be empty")
 	}
 
+	normalizedServiceTier := normalizeServiceTier(serviceTier)
+
 	u.mutex.Lock()
 	defer u.mutex.Unlock()
 
@@ -361,7 +833,11 @@ func (u *AggregatedUsage) AddUsage(model string, inputTokens, outputTokens, cach
 
 	var combo *CostCombination
 	for i := range u.Combinations {
-		if u.Combinations[i].Model == model {
+		comboServiceTier := normalizeServiceTier(u.Combinations[i].ServiceTier)
+		if u.Combinations[i].ServiceTier != comboServiceTier {
+			u.Combinations[i].ServiceTier = comboServiceTier
+		}
+		if u.Combinations[i].Model == model && comboServiceTier == normalizedServiceTier {
 			combo = &u.Combinations[i]
 			break
 		}
@@ -370,6 +846,7 @@ func (u *AggregatedUsage) AddUsage(model string, inputTokens, outputTokens, cach
 	if combo == nil {
 		newCombo := CostCombination{
 			Model:       model,
+			ServiceTier: normalizedServiceTier,
 			Total:       UsageStats{},
 			ByUser:      make(map[string]UsageStats),
 		}