diff --git a/devlog/_plan/260818_fastwire_b2_xai/evidence/010_logs_priority_lower_bound.png b/devlog/_plan/260818_fastwire_b2_xai/evidence/010_logs_priority_lower_bound.png
new file mode 100644
index 0000000000..e7ac2ea007
Binary files /dev/null and b/devlog/_plan/260818_fastwire_b2_xai/evidence/010_logs_priority_lower_bound.png differ
diff --git a/devlog/_plan/260818_fastwire_b2_xai/evidence/README.md b/devlog/_plan/260818_fastwire_b2_xai/evidence/README.md
new file mode 100644
index 0000000000..0ae66d31f3
--- /dev/null
+++ b/devlog/_plan/260818_fastwire_b2_xai/evidence/README.md
@@ -0,0 +1,17 @@
+# FastWire B2 (xAI) — UI evidence
+
+`010_logs_priority_lower_bound.png` — Logs & Debug, three seeded `xai/grok-4.6` rows
+that exercise every branch of the new pricing path:
+
+| Row | Situation | Cost cell |
+| --- | --- | --- |
+| `req-standard` | no Fast requested | `~$0.0300` |
+| `req-priority` | response-confirmed priority, prompt under the long-context threshold | `~$0.0600` — exactly the documented 2x premium over the row above |
+| `req-longctx-priority` | response-confirmed priority, prompt at or above 200k | `≥$0.8760` — the published long-context rate, marked a lower bound because xAI publishes no combined price |
+
+The `≥` prefix is the visible change: a cost that is a known floor rather than an
+estimate now says so instead of rendering as `~$`. The detail drawer explains why
+via the `priority_lower_bound` estimate reason.
+
+Captured against a local proxy with a seeded `usage.jsonl`; no live xAI request was
+billed to produce it.
diff --git a/docs-site/src/content/docs/fr/guides/providers.md b/docs-site/src/content/docs/fr/guides/providers.md
index 59381edd60..40ee8021f1 100644
--- a/docs-site/src/content/docs/fr/guides/providers.md
+++ b/docs-site/src/content/docs/fr/guides/providers.md
@@ -115,7 +115,7 @@ ocx logout <provider>
 
 | Fournisseur | Adaptateur | URL de base | Remarques |
 | --- | --- | --- | --- |
-| `xai` | `openai-chat` | `https://api.x.ai/v1` | Catalogue Grok découvert en direct en priorité ; `grok-4.5` est le modèle de repli par défaut. |
+| `xai` | `openai-chat` | `https://cli-chat-proxy.grok.com/v1` | OAuth utilise la passerelle d'abonnement Grok CLI distincte. Le remplacement par clé API utilise `https://api.x.ai/v1` et peut injecter Priority Processing. Catalogue Grok découvert en direct en priorité ; `grok-4.5` est le modèle de repli par défaut. |
 | `anthropic` | `anthropic` | `https://api.anthropic.com` | Modèles Claude ; liste des modèles récupérée en direct depuis `/v1/models`. |
 | `kimi` | `openai-chat` | `https://api.kimi.com/coding/v1` | Modèles de programmation Kimi K2.7/K2.6/K2.5. |
 | `nous` | `openai-chat` | `https://inference-api.nousresearch.com/v1` | Passerelle d'abonnement Nous Research (le même service en amont que celui utilisé par Hermes Agent). Connexion par autorisation d'appareil auprès de `portal.nousresearch.com` ; le jeton d'accès est le JWT d'inférence envoyé avec chaque requête. Le catalogue mixte de modèles payants et `:free` (`tencent/hy3:free`, `stepfun/step-3.7-flash:free`, ...) est découvert en direct pour le compte connecté. Les jetons d'actualisation sont à usage unique et renouvelés à chaque actualisation. |
diff --git a/docs-site/src/content/docs/guides/providers.md b/docs-site/src/content/docs/guides/providers.md
index 0878773ae9..e58af8360c 100644
--- a/docs-site/src/content/docs/guides/providers.md
+++ b/docs-site/src/content/docs/guides/providers.md
@@ -110,7 +110,7 @@ ocx logout <provider>
 
 | Provider | Adapter | Base URL | Notes |
 | --- | --- | --- | --- |
-| `xai` | `openai-chat` | `https://api.x.ai/v1` | Live-first Grok catalog; `grok-4.5` is the fallback default. |
+| `xai` | `openai-chat` | `https://cli-chat-proxy.grok.com/v1` | OAuth uses the separate Grok CLI subscription gateway. The API-key override uses `https://api.x.ai/v1` and may inject Priority Processing. Live-first Grok catalog; `grok-4.5` is the fallback default. |
 | `anthropic` | `anthropic` | `https://api.anthropic.com` | Claude models; live model list fetched from `/v1/models`. |
 | `kimi` | `openai-chat` | `https://api.kimi.com/coding/v1` | Kimi K2.7/K2.6/K2.5 coding models. |
 | `nous` | `openai-chat` | `https://inference-api.nousresearch.com/v1` | Nous Research subscription gateway (same backend Hermes Agent uses). Device-grant login against `portal.nousresearch.com`; the access token is the per-request inference JWT. Mixed paid + `:free` model catalog (`tencent/hy3:free`, `stepfun/step-3.7-flash:free`, ...) discovered live from the signed-in account. Refresh tokens are single-use and rotated on every refresh. |
diff --git a/docs-site/src/content/docs/ja/guides/providers.md b/docs-site/src/content/docs/ja/guides/providers.md
index 81a19e5e5b..f99bf35aec 100644
--- a/docs-site/src/content/docs/ja/guides/providers.md
+++ b/docs-site/src/content/docs/ja/guides/providers.md
@@ -105,7 +105,7 @@ ocx logout <provider>
 
 | プロバイダー | アダプター | ベース URL | 備考 |
 | --- | --- | --- | --- |
-| `xai` | `openai-chat` | `https://api.x.ai/v1` | ライブ一覧を優先し、フォールバックのデフォルトモデルは `grok-4.5`。 |
+| `xai` | `openai-chat` | `https://cli-chat-proxy.grok.com/v1` | OAuth は独立した Grok CLI サブスクリプションゲートウェイを使用します。API キーのオーバーライドは `https://api.x.ai/v1` を使用し、Priority Processing を注入する場合があります。ライブ一覧を優先し、フォールバックのデフォルトモデルは `grok-4.5`。 |
 | `anthropic` | `anthropic` | `https://api.anthropic.com` | Claude モデル; ライブモデル一覧は `/v1/models` から取得。 |
 | `kimi` | `openai-chat` | `https://api.kimi.com/coding/v1` | Kimi K2.7/K2.6/K2.5 コーディングモデル。 |
 | `nous` | `openai-chat` | `https://inference-api.nousresearch.com/v1` | Nous Research サブスクリプションゲートウェイ（Hermes Agent と同じバックエンド）。`portal.nousresearch.com` へのデバイスグラントログイン; access トークンはリクエストごとの inference JWT。有料 + `:free` モデルの混在カタログ（`tencent/hy3:free`、`stepfun/step-3.7-flash:free` など）はサインイン中のアカウントからライブ探索されます。Refresh トークンは単回使用で、更新のたびにローテーションされます。 |
diff --git a/docs-site/src/content/docs/ko/guides/providers.md b/docs-site/src/content/docs/ko/guides/providers.md
index 4f57dab7cc..b24e17f3fc 100644
--- a/docs-site/src/content/docs/ko/guides/providers.md
+++ b/docs-site/src/content/docs/ko/guides/providers.md
@@ -104,7 +104,7 @@ ocx logout <provider>
 
 | 프로바이더 | 어댑터 | 베이스 URL | 비고 |
 | --- | --- | --- | --- |
-| `xai` | `openai-chat` | `https://api.x.ai/v1` | 실시간 목록을 우선 사용하며, 폴백 기본 모델은 `grok-4.5`입니다. |
+| `xai` | `openai-chat` | `https://cli-chat-proxy.grok.com/v1` | OAuth는 별도의 Grok CLI 구독 게이트웨이를 사용합니다. API 키 오버라이드는 `https://api.x.ai/v1`을 사용하며 Priority Processing을 주입할 수 있습니다. 실시간 목록을 우선 사용하며, 폴백 기본 모델은 `grok-4.5`입니다. |
 | `anthropic` | `anthropic` | `https://api.anthropic.com` | Claude 모델; 실시간 모델 목록은 `/v1/models`에서 가져옵니다. |
 | `kimi` | `openai-chat` | `https://api.kimi.com/coding/v1` | Kimi K2.7/K2.6/K2.5 코딩 모델. |
 | `nous` | `openai-chat` | `https://inference-api.nousresearch.com/v1` | Nous Research 구독 게이트웨이(Hermes Agent와 동일한 백엔드). `portal.nousresearch.com`에 대한 디바이스 그랜트 로그인; access 토큰은 요청별 inference JWT. 유료 + `:free` 모델 혼합 카탈로그(`tencent/hy3:free`, `stepfun/step-3.7-flash:free` 등)는 로그인한 계정에서 실시간으로 발견됩니다. Refresh 토큰은 단회 사용이며, 갱신할 때마다 회전됩니다. |
diff --git a/docs-site/src/content/docs/reference/configuration/providers.md b/docs-site/src/content/docs/reference/configuration/providers.md
index 842d9a6efa..c44b628714 100644
--- a/docs-site/src/content/docs/reference/configuration/providers.md
+++ b/docs-site/src/content/docs/reference/configuration/providers.md
@@ -87,7 +87,7 @@ differing backup and rewrites known legacy namespaced selected ids to bare ids.
 | `modelMaxInputTokens?` | `Record<string, number>` | Positive per-model max input limits used for catalog auto-compaction hints. |
 | `defaultMaxOutputTokens?` | `number` | Provider-wide `openai-chat` fallback when the client omits `max_output_tokens`. |
 | `modelMaxOutputTokens?` | `Record<string, number>` | Positive per-model `openai-chat` fallback budgets; exact/pattern matches beat the provider default. |
-| `modelCosts?` | `Record<string, Cost4>` | Per-model display prices (USD per 1M tokens), keyed by that provider's exact upstream model id — not a provider identifier or a routed `provider/model` label, e.g. `{ "deepseek-v4-flash": { "input": 0.14, "output": 0.28, "cacheRead": 0.0028, "cacheWrite": 0 } }`. Any model id is a valid key — custom providers may target any OpenAI-compatible endpoint through the `openai-chat` adapter, and local or internal provider ids work even when they are absent from the built-in catalogs. User-configured prices win over the built-in catalogs in the Logs `~$` and Usage estimates; historical entries are repriced from the current overlay, so editing a price can move past totals. The fallback order is user `modelCosts` → jawcode catalog → expected-price overlay → model-level vendor fallback, and an all-zero entry falls through to the next source in that sequence. Each rate must be a non-negative finite number at most 1,000,000 (USD per 1M tokens); out-of-range rows are rejected by the management boundary and dropped on load. Display-time estimation only: overlays never affect routing, account selection, quotas, or billing. |
+| `modelCosts?` | `Record<string, Cost4>` | Per-model display prices (USD per 1M tokens), keyed by that provider's exact upstream model id — not a provider identifier or a routed `provider/model` label, e.g. `{ "deepseek-v4-flash": { "input": 0.14, "output": 0.28, "cacheRead": 0.0028, "cacheWrite": 0 } }`. Any model id is a valid key — custom providers may target any OpenAI-compatible endpoint through the `openai-chat` adapter, and local or internal provider ids work even when they are absent from the built-in catalogs. User-configured prices win over the built-in catalogs in the Logs `~$` and Usage estimates; historical entries are repriced from the current overlay, so editing a price can move past totals. The fallback order is user `modelCosts` → exact official correction → jawcode catalog → expected-price overlay → model-level vendor fallback, and an all-zero entry falls through to the next source in that sequence. Each rate must be a non-negative finite number at most 1,000,000 (USD per 1M tokens); out-of-range rows are rejected by the management boundary and dropped on load. Display-time estimation only: overlays never affect routing, account selection, quotas, or billing. |
 | `headers?` | `Record<string, string>` | Extra upstream headers. Authorization, cookies, API-key headers, embedded newlines, and invalid names are rejected. |
 | `openRouterRouting?` | `OpenRouterProviderRouting` | Default OpenRouter `order`, `only`, and `allowFallbacks` preferences; valid only for canonical OpenRouter with `openai-chat`. |
 | `modelOpenRouterRouting?` | `Record<string, OpenRouterProviderRouting>` | Exact model-id overrides that replace the provider-wide OpenRouter preference. |
@@ -152,6 +152,25 @@ contract; existing configurations see these migration deltas:
 
 Explicit capability `false` and Responses caller-tier forwarding retain their existing contracts.
 
+### xAI Priority Processing
+
+The built-in `xai` preset advertises and injects Fast only when its effective transport uses
+`authMode: "key"`. API-key mode targets `https://api.x.ai/v1` through the `openai-chat` adapter and
+sends `service_tier: "priority"` through Chat Completions. `ocx login xai`
+instead stores OAuth credentials for the separate Grok CLI subscription-gateway flow, so OAuth
+remains unclassified: its catalog rows do not advertise Fast and the proxy does not inject a tier.
+
+xAI charges Priority Processing at 2× the standard token price for input, output, cached, and
+reasoning tokens; cache discounts are applied before the multiplier. Cost estimates use that premium
+only when xAI's response confirms `service_tier: "priority"`. A missing or unparsed response tier is
+not confirmation, and an echoed `default` is a downgrade; all three stay at the standard price.
+
+For `grok-4.6`, the standard rate per 1M tokens is $2.00 input, $0.50 cached input, and $6.00
+output. A prompt of at least 200,000 tokens reprices the whole request at $4.00 / $1.00 / $12.00.
+xAI has not published how that long-context band combines with Priority Processing. When a
+long-context response confirms `priority`, the dashboard therefore shows the published long-context
+cost with a `≥` marker and a lower-bound explanation; it never invents a stacked multiplier.
+
 ### OpenRouter Fast
 
 The canonical `https://openrouter.ai/api/v1` preset advertises Fast only for these exact
diff --git a/docs-site/src/content/docs/ru/guides/providers.md b/docs-site/src/content/docs/ru/guides/providers.md
index 1966d9db63..1dfe171a58 100644
--- a/docs-site/src/content/docs/ru/guides/providers.md
+++ b/docs-site/src/content/docs/ru/guides/providers.md
@@ -114,7 +114,7 @@ ocx logout <provider>
 
 | Провайдер | Адаптер | Базовый URL | Примечания |
 | --- | --- | --- | --- |
-| `xai` | `openai-chat` | `https://api.x.ai/v1` | Каталог Grok загружается в реальном времени; фолбэк по умолчанию — `grok-4.5`. |
+| `xai` | `openai-chat` | `https://cli-chat-proxy.grok.com/v1` | OAuth использует отдельный шлюз подписки Grok CLI. Переопределение с API-ключом использует `https://api.x.ai/v1` и может добавлять Priority Processing. Каталог Grok загружается в реальном времени; фолбэк по умолчанию — `grok-4.5`. |
 | `anthropic` | `anthropic` | `https://api.anthropic.com` | Модели Claude; актуальный список моделей загружается из `/v1/models`. |
 | `kimi` | `openai-chat` | `https://api.kimi.com/coding/v1` | Модели Kimi K2.7/K2.6/K2.5 для кодинга. |
 | `nous` | `openai-chat` | `https://inference-api.nousresearch.com/v1` | Шлюз подписки Nous Research (тот же бэкенд, что использует Hermes Agent). Вход по device grant против `portal.nousresearch.com`; access-токен — это JWT для каждого запроса к inference. Смешанный каталог платных + `:free` моделей (`tencent/hy3:free`, `stepfun/step-3.7-flash:free`, …) обнаруживается вживую по авторизованному аккаунту. Refresh-токены одноразовые и ротируются при каждом обновлении. |
diff --git a/docs-site/src/content/docs/tr/guides/providers.md b/docs-site/src/content/docs/tr/guides/providers.md
index ee153a0780..15e2ab3cf4 100644
--- a/docs-site/src/content/docs/tr/guides/providers.md
+++ b/docs-site/src/content/docs/tr/guides/providers.md
@@ -129,7 +129,7 @@ ocx logout <saglayici>
 
 | Sağlayıcı | Adaptör | Temel URL | Notlar |
 | --- | --- | --- | --- |
-| `xai` | `openai-chat` | `https://api.x.ai/v1` | Canlı öncelikli Grok kataloğu; `grok-4.5` geri dönüş varsayılanıdır. |
+| `xai` | `openai-chat` | `https://cli-chat-proxy.grok.com/v1` | OAuth ayrı Grok CLI abonelik ağ geçidini kullanır. API anahtarı geçersiz kılması `https://api.x.ai/v1` kullanır ve Priority Processing ekleyebilir. Canlı öncelikli Grok kataloğu; `grok-4.5` geri dönüş varsayılanıdır. |
 | `anthropic` | `anthropic` | `https://api.anthropic.com` | Claude modelleri; canlı model listesi `/v1/models` üzerinden getirilir. |
 | `kimi` | `openai-chat` | `https://api.kimi.com/coding/v1` | Kimi K2.7/K2.6/K2.5 kodlama modelleri. |
 | `nous` | `openai-chat` | `https://inference-api.nousresearch.com/v1` | Nous Research abonelik ağ geçidi (Hermes Agent'ın kullandığı aynı arka uç). `portal.nousresearch.com`'a karşı cihaz yetkilendirmesi girişi; erişim belirteci istek başına çıkarım JWT'sidir. Oturum açmış hesaptan canlı olarak keşfedilen karışık ücretli + `:free` model kataloğu (`tencent/hy3:free`, `stepfun/step-3.7-flash:free`, ...). Yenileme belirteçleri tek kullanımlıktır ve her yenilemede döndürülür. |
diff --git a/docs-site/src/content/docs/zh-cn/guides/providers.md b/docs-site/src/content/docs/zh-cn/guides/providers.md
index 4e924458ee..a73676e496 100644
--- a/docs-site/src/content/docs/zh-cn/guides/providers.md
+++ b/docs-site/src/content/docs/zh-cn/guides/providers.md
@@ -95,7 +95,7 @@ ocx logout <provider>
 
 | 提供商 | Adapter | 基础 URL | 备注 |
 | --- | --- | --- | --- |
-| `xai` | `openai-chat` | `https://api.x.ai/v1` | 优先使用实时 Grok 目录；回退默认模型为 `grok-4.5`。 |
+| `xai` | `openai-chat` | `https://cli-chat-proxy.grok.com/v1` | OAuth 使用独立的 Grok CLI 订阅网关。API 密钥覆盖模式使用 `https://api.x.ai/v1`，并可能注入 Priority Processing。优先使用实时 Grok 目录；回退默认模型为 `grok-4.5`。 |
 | `anthropic` | `anthropic` | `https://api.anthropic.com` | Claude 模型；实时模型列表从 `/v1/models` 获取。 |
 | `kimi` | `openai-chat` | `https://api.kimi.com/coding/v1` | Kimi K2.7/K2.6/K2.5 编程模型。 |
 | `nous` | `openai-chat` | `https://inference-api.nousresearch.com/v1` | Nous Research 订阅网关（与 Hermes Agent 使用同一后端）。通过设备授权登录 `portal.nousresearch.com`；access 令牌是每个请求的 inference JWT。付费 + `:free` 模型混合目录（`tencent/hy3:free`、`stepfun/step-3.7-flash:free` 等）会从已登录账户实时发现。Refresh 令牌是单次使用，每次刷新都会轮换。 |
diff --git a/docs-site/src/content/docs/zh-tw/guides/providers.md b/docs-site/src/content/docs/zh-tw/guides/providers.md
index ee7d709880..28298c768a 100644
--- a/docs-site/src/content/docs/zh-tw/guides/providers.md
+++ b/docs-site/src/content/docs/zh-tw/guides/providers.md
@@ -103,7 +103,7 @@ ocx logout <provider>
 
 | 供應商 | Adapter | Base URL | 備註 |
 | --- | --- | --- | --- |
-| `xai` | `openai-chat` | `https://api.x.ai/v1` | 優先使用即時 Grok catalog；fallback 預設為 `grok-4.5`。 |
+| `xai` | `openai-chat` | `https://cli-chat-proxy.grok.com/v1` | OAuth 使用獨立的 Grok CLI 訂閱 gateway。API key 覆寫使用 `https://api.x.ai/v1`，並可能注入 Priority Processing。優先使用即時 Grok catalog；fallback 預設為 `grok-4.5`。 |
 | `anthropic` | `anthropic` | `https://api.anthropic.com` | Claude 模型；即時模型列表從 `/v1/models` 取得。 |
 | `kimi` | `openai-chat` | `https://api.kimi.com/coding/v1` | Kimi K2.7/K2.6/K2.5 coding 模型。 |
 | `nous` | `openai-chat` | `https://inference-api.nousresearch.com/v1` | Nous Research 訂閱 gateway（Hermes Agent 使用相同 backend）。透過 `portal.nousresearch.com` 做 device-grant 登入；access token 是每次請求使用的 inference JWT。混合付費與 `:free` 模型 catalog（`tencent/hy3:free`、`stepfun/step-3.7-flash:free` 等）會從已登入帳號即時探索。Refresh token 為單次使用，每次 refresh 都會輪換。 |
diff --git a/gui/src/i18n/de.ts b/gui/src/i18n/de.ts
index 8e64e29b97..41ae5d02e5 100644
--- a/gui/src/i18n/de.ts
+++ b/gui/src/i18n/de.ts
@@ -717,7 +717,7 @@ export const de: Record<TKey, string> = {
   "logs.detail.estimate.cache_detail_missing": "Cache-Details fehlen; Eingabe ist als Obergrenze geschätzt.",
   "logs.detail.estimate.expected_price_overlay": "Ein verifizierter Expected-Listenpreis wurde verwendet.",
   "logs.detail.estimate.provider_cost_overlay": "Ein vom Anbieter konfiguriertes Preis-Overlay wurde verwendet.",
-  "logs.detail.estimate.priority_lower_bound": "Der bestätigte OpenRouter-Priority-Preis ist nicht verfügbar; die angezeigte Standardpreisschätzung ist eine bekannte Untergrenze.",
+  "logs.detail.estimate.priority_lower_bound": "Der bestätigte Priority-Preis ist nicht verfügbar; die angezeigte Schätzung ist eine bekannte Untergrenze.",
   "logs.col.error": "Fehler",
   "logs.col.upstreamReason": "Upstream-Grund",
   "logs.col.duration": "Dauer",
diff --git a/gui/src/i18n/en.ts b/gui/src/i18n/en.ts
index 262e913162..01d69f0fd1 100644
--- a/gui/src/i18n/en.ts
+++ b/gui/src/i18n/en.ts
@@ -750,7 +750,7 @@ export const en = {
   "logs.detail.estimate.cache_detail_missing": "Cache details were unavailable; input is an upper-bound estimate.",
   "logs.detail.estimate.expected_price_overlay": "A verified expected list price was used.",
   "logs.detail.estimate.provider_cost_overlay": "A provider-configured price overlay was used.",
-  "logs.detail.estimate.priority_lower_bound": "The confirmed OpenRouter priority price is unavailable; the displayed standard-price estimate is a known lower bound.",
+  "logs.detail.estimate.priority_lower_bound": "The confirmed Priority price is unavailable; the displayed estimate is a known lower bound.",
   "logs.col.error": "Error",
   "logs.col.upstreamReason": "Upstream reason",
   "logs.col.duration": "Duration",
diff --git a/gui/src/i18n/fr.ts b/gui/src/i18n/fr.ts
index 147773fd93..3765f4d95f 100644
--- a/gui/src/i18n/fr.ts
+++ b/gui/src/i18n/fr.ts
@@ -731,7 +731,7 @@ export const fr: Record<TKey, string> = {
   "logs.detail.estimate.cache_detail_missing": "Les détails du cache n’étaient pas disponibles ; l’entrée est une estimation de la limite supérieure.",
   "logs.detail.estimate.expected_price_overlay": "Un tarif catalogue attendu et vérifié a été utilisé.",
   "logs.detail.estimate.provider_cost_overlay": "Un remplacement de tarif configuré pour le fournisseur a été utilisé.",
-  "logs.detail.estimate.priority_lower_bound": "Le tarif Priority OpenRouter confirmé n’est pas disponible ; l’estimation au tarif standard affichée est une borne inférieure connue.",
+  "logs.detail.estimate.priority_lower_bound": "Le tarif Priority confirmé n’est pas disponible ; l’estimation affichée est une borne inférieure connue.",
   "logs.col.error": "Erreur",
   "logs.col.upstreamReason": "Motif en amont",
   "logs.col.duration": "Durée",
diff --git a/gui/src/i18n/ja.ts b/gui/src/i18n/ja.ts
index 11a85127ea..63e20efd3b 100644
--- a/gui/src/i18n/ja.ts
+++ b/gui/src/i18n/ja.ts
@@ -693,7 +693,7 @@ export const ja: Record<TKey, string> = {
   "logs.detail.estimate.cache_detail_missing": "キャッシュの詳細が利用できませんでした; 入力は上限の推定です。",
   "logs.detail.estimate.expected_price_overlay": "検証済みの予想定価が使用されました。",
   "logs.detail.estimate.provider_cost_overlay": "プロバイダー設定の価格オーバーレイが使用されました。",
-  "logs.detail.estimate.priority_lower_bound": "確認済みの OpenRouter Priority 価格は取得できないため、表示される標準価格の見積もりは既知の下限です。",
+  "logs.detail.estimate.priority_lower_bound": "確認済みの Priority 価格を利用できないため、表示される見積もりは既知の下限です。",
   "logs.col.error": "エラー",
   "logs.col.upstreamReason": "上流の理由",
   "logs.col.duration": "所要時間",
diff --git a/gui/src/i18n/ko.ts b/gui/src/i18n/ko.ts
index 5bbc14ae7d..b13cd141ff 100644
--- a/gui/src/i18n/ko.ts
+++ b/gui/src/i18n/ko.ts
@@ -736,7 +736,7 @@ export const ko: Record<TKey, string> = {
   "logs.detail.estimate.cache_detail_missing": "캐시 상세가 없어 입력 전액을 상한으로 추정했습니다.",
   "logs.detail.estimate.expected_price_overlay": "검증된 expected 정가를 사용했습니다.",
   "logs.detail.estimate.provider_cost_overlay": "프로바이더 구성 가격 오버레이를 사용했습니다.",
-  "logs.detail.estimate.priority_lower_bound": "확인된 OpenRouter Priority 가격을 사용할 수 없어 표시된 표준 가격 추정치는 알려진 하한입니다.",
+  "logs.detail.estimate.priority_lower_bound": "확인된 Priority 가격을 사용할 수 없어 표시된 추정치는 알려진 하한입니다.",
   "logs.col.error": "오류",
   "logs.col.upstreamReason": "업스트림 원인",
   "logs.col.duration": "소요 시간",
diff --git a/gui/src/i18n/ru.ts b/gui/src/i18n/ru.ts
index 396ccf3ed0..05cf793779 100644
--- a/gui/src/i18n/ru.ts
+++ b/gui/src/i18n/ru.ts
@@ -734,7 +734,7 @@ export const ru: Record<TKey, string> = {
   "logs.detail.estimate.cache_detail_missing": "Детализация кэша недоступна; входные токены оценены по верхней границе.",
   "logs.detail.estimate.expected_price_overlay": "Использована подтверждённая ожидаемая цена из прайс-листа.",
   "logs.detail.estimate.provider_cost_overlay": "Использован ценовой оверлей провайдера.",
-  "logs.detail.estimate.priority_lower_bound": "Подтверждённая цена OpenRouter Priority недоступна; показанная оценка по стандартной цене является известной нижней границей.",
+  "logs.detail.estimate.priority_lower_bound": "Подтверждённая цена Priority недоступна; показанная оценка является известной нижней границей.",
   "logs.col.error": "Ошибка",
   "logs.col.upstreamReason": "Причина от провайдера",
   "logs.col.duration": "Длительность",
diff --git a/gui/src/i18n/tr.ts b/gui/src/i18n/tr.ts
index f460bbeb36..c98421183e 100644
--- a/gui/src/i18n/tr.ts
+++ b/gui/src/i18n/tr.ts
@@ -741,7 +741,7 @@ export const tr: Record<TKey, string> = {
   "logs.detail.estimate.cache_detail_missing": "Önbellek detayları eksik.",
   "logs.detail.estimate.expected_price_overlay": "Doğrulanmış liste fiyatı kullanıldı.",
   "logs.detail.estimate.provider_cost_overlay": "Kullanıcı tarafından yapılandırılan bir sağlayıcı fiyat katmanı kullanıldı.",
-  "logs.detail.estimate.priority_lower_bound": "Doğrulanan OpenRouter Priority fiyatı kullanılamıyor; gösterilen standart fiyat tahmini bilinen bir alt sınırdır.",
+  "logs.detail.estimate.priority_lower_bound": "Doğrulanan Priority fiyatı kullanılamıyor; gösterilen tahmin bilinen bir alt sınırdır.",
   "logs.col.error": "Hata",
   "logs.col.upstreamReason": "Yukarı akış nedeni",
   "logs.col.duration": "Süre",
diff --git a/gui/src/i18n/zh-TW.ts b/gui/src/i18n/zh-TW.ts
index 942e21e61f..d35a96bd34 100644
--- a/gui/src/i18n/zh-TW.ts
+++ b/gui/src/i18n/zh-TW.ts
@@ -1818,7 +1818,7 @@ export const zhTW: Record<TKey, string> = {
   "logs.detail.attempt.recovery.emptyCompletion": "空白完成重試",
   "logs.detail.attempt.recovery.unknown": "未知的復原原因",
   "logs.detail.estimate.provider_cost_overlay": "已使用供應商設定的價格覆蓋。",
-  "logs.detail.estimate.priority_lower_bound": "無法取得已確認的 OpenRouter Priority 價格；目前顯示的標準價格估算是已知下限。",
+  "logs.detail.estimate.priority_lower_bound": "無法取得已確認的 Priority 價格；目前顯示的估算是已知下限。",
   "pws.cockpitImportDescription": "從此裝置匯入 Cockpit Tools Antigravity JSON 匯出檔。不會顯示檔案內容。",
   "pws.cockpitImportFileLabel": "Cockpit Tools Antigravity JSON 匯出檔",
   "pws.cockpitImportChooseFile": "選擇 JSON 檔案",
diff --git a/gui/src/i18n/zh.ts b/gui/src/i18n/zh.ts
index b9cd4a3573..dc749ac530 100644
--- a/gui/src/i18n/zh.ts
+++ b/gui/src/i18n/zh.ts
@@ -729,7 +729,7 @@ export const zh: Record<TKey, string> = {
   "logs.detail.estimate.cache_detail_missing": "缺少缓存明细；输入费用按上限估算。",
   "logs.detail.estimate.expected_price_overlay": "使用了已验证的 Expected 标价。",
   "logs.detail.estimate.provider_cost_overlay": "使用了用户配置的提供方价格覆盖。",
-  "logs.detail.estimate.priority_lower_bound": "暂无已确认的 OpenRouter Priority 价格；当前显示的标准价估算是已知下界。",
+  "logs.detail.estimate.priority_lower_bound": "暂无已确认的 Priority 价格；当前显示的估算是已知下界。",
   "logs.col.error": "错误",
   "logs.col.upstreamReason": "上游原因",
   "logs.col.duration": "耗时",
diff --git a/gui/tests/logs-cost-lower-bound.test.ts b/gui/tests/logs-cost-lower-bound.test.ts
new file mode 100644
index 0000000000..60759219fa
--- /dev/null
+++ b/gui/tests/logs-cost-lower-bound.test.ts
@@ -0,0 +1,60 @@
+import { describe, expect, test } from "bun:test";
+import { DICTS } from "../src/i18n/catalogs";
+import { interpolate, type Locale, type TFn } from "../src/i18n/shared";
+import {
+  formatEstimatedUsd,
+  formatEstimatedUsdValue,
+  summarizeEstimatedCosts,
+} from "../src/pages/logs-cost-format";
+
+function translator(locale: Locale): TFn {
+  return (key, vars) => interpolate(DICTS[locale][key], vars);
+}
+
+test("ordinary dashboard costs retain the estimate marker", () => {
+  expect(formatEstimatedUsdValue(0.77, translator("en"), "en-US", false)).toBe("~$0.7700");
+});
+
+test("priority long-context lower bounds render with a greater-than-or-equal marker", () => {
+  expect(formatEstimatedUsdValue(0.77, translator("en"), "en-US", true)).toBe("≥$0.7700");
+});
+
+test("USD placement and separators follow a non-English locale", () => {
+  expect(formatEstimatedUsdValue(0.77, translator("de"), "de-DE", false)).toBe("ca. 0,7700\u00a0$");
+  expect(formatEstimatedUsd({ kind: "unavailable" }, translator("de"), "de-DE")).toBe("nicht verfügbar");
+});
+
+describe("conversation cost lower-bound aggregation", () => {
+  const priced = (total: number, lowerBound: boolean) => ({
+    usageStatus: "reported",
+    displayMetrics: {
+      cost: {
+        kind: "value" as const,
+        estimate: { cost: { total }, priorityLowerBound: lowerBound },
+      },
+    },
+  });
+
+  test("marks a total only when every included priced estimate is a lower bound", () => {
+    expect(summarizeEstimatedCosts([priced(0.77, true), priced(1.23, true)])).toMatchObject({
+      estimatedCostUsd: 2,
+      priorityLowerBound: true,
+    });
+    expect(summarizeEstimatedCosts([priced(0.77, true), priced(1.23, false)])).toMatchObject({
+      estimatedCostUsd: 2,
+      priorityLowerBound: false,
+    });
+  });
+
+  test("preserves unpriced and unsupported exclusions without minting a lower bound", () => {
+    expect(summarizeEstimatedCosts([
+      { usageStatus: "reported", displayMetrics: { cost: { kind: "unavailable" } } },
+      { usageStatus: "unsupported" },
+    ])).toEqual({
+      estimatedCostUsd: 0,
+      priorityLowerBound: false,
+      unpricedRequests: 1,
+      unmeteredRequests: 1,
+    });
+  });
+});
diff --git a/src/codex/catalog/effort.ts b/src/codex/catalog/effort.ts
index 3bf5daa086..0648b64d17 100644
--- a/src/codex/catalog/effort.ts
+++ b/src/codex/catalog/effort.ts
@@ -147,7 +147,7 @@ export function applyCatalogModelMetadata(entry: RawEntry, model?: CatalogModel)
     entry.service_tiers = [{
       id: "priority",
       name: "Fast",
-      description: "1.5x speed, increased usage",
+      description: model.fastTierDescription ?? "1.5x speed, increased usage",
     }];
     entry.additional_speed_tiers = ["fast"];
   }
diff --git a/src/codex/catalog/parsing.ts b/src/codex/catalog/parsing.ts
index 0d1b2c2aaa..a2a1c86c7c 100644
--- a/src/codex/catalog/parsing.ts
+++ b/src/codex/catalog/parsing.ts
@@ -125,6 +125,8 @@ export interface CatalogModel {
   supportsVerbosity?: boolean;
   /** Whether this exact routed model has a verified OpenAI-compatible service tier. */
   supportsServiceTier?: boolean;
+  /** Optional provider-specific copy for the advertised Fast tier. */
+  fastTierDescription?: string;
   supportsReasoningSummaries?: boolean;
   /**
    * Codex tool calling mode for this routed model.
diff --git a/src/codex/catalog/provider-fetch.ts b/src/codex/catalog/provider-fetch.ts
index 9c3384b4e3..565f6057a4 100644
--- a/src/codex/catalog/provider-fetch.ts
+++ b/src/codex/catalog/provider-fetch.ts
@@ -34,7 +34,8 @@ import { getModelMetadata, getModelMetadataCaseInsensitive, listModelMetadata, r
 import { enrichProviderFromRegistry, shouldCaseFoldMetadataModelId } from "../../providers/derive";
 import {
   captureFastPolicyAuthority,
-  serviceTierSupportForModel,
+  fastPolicyForModel,
+  serviceTierSupportFromPolicy,
 } from "../../providers/service-tier";
 import type { FastPolicyAuthority } from "../../providers/fastwire";
 import { effectiveGoogleMode, getProviderRegistryEntry, providerMatchesRegistryTransport } from "../../providers/registry";
@@ -647,8 +648,13 @@ export function applyProviderConfigHints(name: string, prov: OcxProviderConfig,
   const reasoningEfforts = configuredReasoningEfforts(prov, model.id);
   const defaultReasoningEffort = modelRecordValue(prov.modelDefaultReasoningEfforts, model.id) ?? model.defaultReasoningEffort;
   const supportsReasoningSummaries = configuredReasoningSummarySupport(prov, model.id);
-  const supportsServiceTier = serviceTierSupportForModel(prov, model.id, name);
-  const { supportsServiceTier: _staleServiceTier, ...modelWithoutServiceTier } = model;
+  const fastPolicy = fastPolicyForModel(prov, model.id, name);
+  const supportsServiceTier = serviceTierSupportFromPolicy(fastPolicy);
+  const {
+    supportsServiceTier: _staleServiceTier,
+    fastTierDescription: _staleFastTierDescription,
+    ...modelWithoutServiceTier
+  } = model;
   // 已发现窗口只允许被配置值压低；缺窗口时，已开的 Context cap 就是实际窗口。
   const discoveredWindow = typeof model.contextWindow === "number" && model.contextWindow > 0
     ? model.contextWindow
@@ -671,6 +677,9 @@ export function applyProviderConfigHints(name: string, prov: OcxProviderConfig,
     ...(defaultReasoningEffort ? { defaultReasoningEffort } : {}),
     ...(typeof supportsReasoningSummaries === "boolean" ? { supportsReasoningSummaries } : {}),
     ...(typeof supportsServiceTier === "boolean" ? { supportsServiceTier } : {}),
+    ...(supportsServiceTier === true && fastPolicy.fastTierDescription !== undefined
+      ? { fastTierDescription: fastPolicy.fastTierDescription }
+      : {}),
     ...(prov.adapter === "kiro" ? { supportsVerbosity: false } : {}),
     // Default-on for openai-chat providers (explicit false opts out); other adapters
     // advertise only on explicit opt-in.
@@ -1845,8 +1854,11 @@ async function gatherRoutedModelsUncached(
       ? nativeDefaultReasoningEffort(cm.modelId)
       : undefined;
     const supportsReasoningSummaries = configuredReasoningSummarySupport(rawProvider, cm.modelId);
-    const supportsServiceTier = effectiveProvider
-      ? serviceTierSupportForModel(effectiveProvider, cm.modelId, cm.provider)
+    const fastPolicy = effectiveProvider
+      ? fastPolicyForModel(effectiveProvider, cm.modelId, cm.provider)
+      : undefined;
+    const supportsServiceTier = fastPolicy
+      ? serviceTierSupportFromPolicy(fastPolicy)
       : undefined;
     const base: CatalogModel = {
       id: cm.modelId,
@@ -1883,6 +1895,9 @@ async function gatherRoutedModelsUncached(
       ...(Array.isArray(cm.reasoningEfforts) ? { reasoningEfforts: [...cm.reasoningEfforts] } : {}),
       ...(cm.defaultReasoningEffort ? { defaultReasoningEffort: cm.defaultReasoningEffort } : {}),
       ...(typeof supportsServiceTier === "boolean" ? { supportsServiceTier } : {}),
+      ...(supportsServiceTier === true && fastPolicy?.fastTierDescription !== undefined
+        ? { fastTierDescription: fastPolicy.fastTierDescription }
+        : {}),
       ...(cm.codexToolMode !== undefined
         ? { codexToolMode: cm.codexToolMode }
         : effectiveProvider?.codexToolMode !== undefined
diff --git a/src/providers/fastwire.ts b/src/providers/fastwire.ts
index 63cd642532..9098bd9df9 100644
--- a/src/providers/fastwire.ts
+++ b/src/providers/fastwire.ts
@@ -33,6 +33,7 @@ export interface FastPolicyAuthority {
   readonly providerAdapter: string;
   readonly providerAuthMode?: ProviderAuthKind;
   readonly fastWireDeclaration: FastWire | null | undefined;
+  readonly fastTierDescription?: string;
   readonly modelWireOverrideAllowed: boolean;
   readonly authTransport: FastPolicyAuthTransport;
   readonly capability: {
@@ -55,6 +56,7 @@ export interface ResolvedFastPolicy {
     | "pin-unavailable";
   readonly adapter: string;
   readonly fastWire: FastWire | null;
+  readonly fastTierDescription?: string;
   readonly forwardCallerTier: boolean;
 }
 
@@ -222,7 +224,16 @@ export function resolveFastPolicy(
   else if (capability === undefined) eligibility = "unclassified";
   else eligibility = "eligible";
 
-  return { capability, eligibility, adapter, fastWire, forwardCallerTier };
+  return {
+    capability,
+    eligibility,
+    adapter,
+    fastWire,
+    ...(authority.fastTierDescription !== undefined
+      ? { fastTierDescription: authority.fastTierDescription }
+      : {}),
+    forwardCallerTier,
+  };
 }
 
 export function canonicalFastTierMarker(callerTier: string | undefined): "priority" | undefined {
diff --git a/src/providers/registry.ts b/src/providers/registry.ts
index 4e61bb1e68..32a65a7af5 100644
--- a/src/providers/registry.ts
+++ b/src/providers/registry.ts
@@ -227,6 +227,18 @@ export interface ProviderRegistryEntry {
   supportsOpenAiWebSearchToolFields?: boolean;
   /** Registry default for exact model service-tier capability; explicit config keys win. */
   modelSupportsServiceTier?: Record<string, boolean>;
+  /**
+   * Registry-only service-tier defaults for an OAuth preset's explicit API-key transport.
+   * Applied only when `allowKeyAuthOverride` is true and the captured effective auth transport
+   * is key-based. Explicit provider config still wins field-by-field, including `false`.
+   */
+  keyAuthServiceTier?: {
+    supportsServiceTier?: boolean;
+    modelSupportsServiceTier?: Record<string, boolean>;
+    chatServiceTier?: boolean;
+  };
+  /** Provider-specific copy for the Codex catalog's Fast tier. */
+  fastTierDescription?: string;
   /**
    * Registry-only destination guard for `modelSupportsServiceTier`. This scopes vendor evidence
    * without changing provider ownership, routing, authentication, or config validation.
@@ -1023,6 +1035,14 @@ export const PROVIDER_REGISTRY: readonly ProviderRegistryEntry[] = [
     baseUrl: "https://api.x.ai/v1",
     authKind: "oauth",
     allowKeyAuthOverride: true,
+    // Priority Processing is documented for xAI's public API-key Chat Completions and
+    // Responses endpoints. OAuth is a separate Grok CLI subscription gateway and remains
+    // unclassified; do not turn this into a provider-wide supportsServiceTier declaration.
+    keyAuthServiceTier: {
+      supportsServiceTier: true,
+      chatServiceTier: true,
+    },
+    fastTierDescription: "Priority processing, 2x token price",
     featured: true,
     oauthId: "xai",
     jawcodeBundle: "xai",
diff --git a/src/providers/service-tier.ts b/src/providers/service-tier.ts
index 2a09530d23..de71cb3f9e 100644
--- a/src/providers/service-tier.ts
+++ b/src/providers/service-tier.ts
@@ -70,11 +70,22 @@ function buildFastPolicyAuthority(
   capabilityProvider: ServiceTierCapabilityProvider = provider,
 ): FastPolicyAuthority {
   const registry = registryTransportMatch ? getProviderRegistryEntry(providerName) : undefined;
+  const authTransport = resolveProviderAuthTransport(
+    provider.adapter,
+    provider.authMode ?? registry?.authKind ?? "key",
+    provider.apiKeyTransport,
+  );
+  const keyAuthDefaults = registry?.allowKeyAuthOverride === true
+    && (authTransport === "authorization_bearer" || authTransport === "x_api_key")
+    ? registry.keyAuthServiceTier
+    : undefined;
   const registryModelCapabilities = registry
     && registryModelServiceTierCapabilityApplies(registry, capabilityProvider)
     ? registry.modelSupportsServiceTier
     : undefined;
-  const providerCapability = capabilityProvider.supportsServiceTier ?? registry?.supportsServiceTier;
+  const providerCapability = capabilityProvider.supportsServiceTier
+    ?? keyAuthDefaults?.supportsServiceTier
+    ?? registry?.supportsServiceTier;
   const authority: FastPolicyAuthority = Object.freeze({
     providerAdapter: provider.adapter,
     providerAuthMode: provider.authMode ?? registry?.authKind ?? "key",
@@ -82,19 +93,23 @@ function buildFastPolicyAuthority(
       provider.fastWire !== undefined ? provider.fastWire : registry?.fastWire,
       { freeze: true },
     ),
+    ...(registry?.fastTierDescription !== undefined
+      ? { fastTierDescription: registry.fastTierDescription }
+      : {}),
     modelWireOverrideAllowed: !isCanonicalOpenAiForwardProvider(provider as OcxProviderConfig),
-    authTransport: resolveProviderAuthTransport(
-      provider.adapter,
-      provider.authMode ?? registry?.authKind ?? "key",
-      provider.apiKeyTransport,
-    ),
+    authTransport,
     capability: Object.freeze({
       ...(providerCapability !== undefined ? { provider: providerCapability } : {}),
       models: Object.freeze({
         ...(registryModelCapabilities ?? {}),
+        ...(keyAuthDefaults?.modelSupportsServiceTier ?? {}),
         ...(capabilityProvider.modelSupportsServiceTier ?? {}),
       }),
-      ...(provider.chatServiceTier !== undefined ? { chatServiceTier: provider.chatServiceTier } : {}),
+      ...(provider.chatServiceTier !== undefined
+        ? { chatServiceTier: provider.chatServiceTier }
+        : keyAuthDefaults?.chatServiceTier !== undefined
+          ? { chatServiceTier: keyAuthDefaults.chatServiceTier }
+          : {}),
     }),
     modelAdapters: Object.freeze({ ...(provider.modelAdapters ?? {}) }),
     hardPins: captureWireAdapterHardPins(providerName),
diff --git a/src/usage/cost.ts b/src/usage/cost.ts
index 608dea31ac..d507668a84 100644
--- a/src/usage/cost.ts
+++ b/src/usage/cost.ts
@@ -22,7 +22,8 @@ import { activeConfiguredProviders, activeUserCostOverlays, userCostOverlayVersi
 import {
   EXPECTED_PRICE_OVERLAYS,
   findExpectedPriceOverlay,
-  resolvePriorityMultiplier,
+  findVerifiedPriceOverride,
+  findPriorityPricingRule,
   findContextTier,
   isLongContext,
   type Cost4,
@@ -81,12 +82,12 @@ export interface AttemptCostEstimate {
   price: MatchedPrice;
   cost: CostBreakdown;
   estimated: boolean;
-  /** Applied OpenAI priority-tier multiplier (undefined or 1 = standard). */
+  /** Applied provider priority-tier multiplier (undefined or 1 = standard). */
   priorityMultiplier?: number;
-  /** Standard-price estimate is a known floor for a confirmed, unpriced priority endpoint. */
-  priorityLowerBound?: boolean;
   /** Set when the published long-context rate was applied (#908). */
   contextTier?: ContextTierName;
+  /** The numeric estimate is a known floor because the exact Priority price is unavailable. */
+  priorityLowerBound?: boolean;
 }
 
 export interface CostEstimate {
@@ -95,12 +96,12 @@ export interface CostEstimate {
   estimated: boolean;
   attempts?: AttemptCostEstimate[];
   price?: MatchedPrice;
-  /** Applied OpenAI priority-tier multiplier (undefined or 1 = standard). */
+  /** Applied provider priority-tier multiplier (undefined or 1 = standard). */
   priorityMultiplier?: number;
-  /** Standard-price estimate is a known floor for a confirmed, unpriced priority endpoint. */
-  priorityLowerBound?: boolean;
   /** Set when any priced attempt used the published long-context rate (#908). */
   contextTier?: ContextTierName;
+  /** The aggregate is a known floor because every priced attempt is a lower bound. */
+  priorityLowerBound?: boolean;
 }
 
 function finiteNonNegative(value: number): boolean {
@@ -236,7 +237,7 @@ function resolveMatchedPriceInner(
 
 /**
  * Exact provider/model price lookup: user-configured `modelCosts` first, then
- * the jawcode provider bundle, then the expected-price overlay, then the
+ * an exact official correction, the jawcode provider bundle, the expected-price overlay, then the
  * model-level vendor fallback. All-zero rows fall through ("not billable").
  */
 function resolveMatchedPriceExact(
@@ -249,6 +250,20 @@ function resolveMatchedPriceExact(
   // operator's explicit price is authoritative for the ~$ estimate.
   const userOverlay = userOverlayMatch(provider, modelId, userOverlays);
   if (userOverlay) return userOverlay;
+  const verifiedOverride = overlays === EXPECTED_PRICE_OVERLAYS
+    ? findVerifiedPriceOverride(provider, modelId)
+    : undefined;
+  if (verifiedOverride && validCost4(verifiedOverride.cost4) && hasNonZeroCost(verifiedOverride.cost4)) {
+    return {
+      provider,
+      modelId,
+      cost4: verifiedOverride.cost4,
+      source: "expected",
+      sourceRef: verifiedOverride.source,
+      verifiedAt: verifiedOverride.verifiedAt,
+      status: "verified",
+    };
+  }
   const metadataProvider = resolveMetadataProvider(provider);
   const bundled = metadataProvider
     ? getModelMetadata(metadataProvider, modelId)
@@ -320,13 +335,6 @@ function isEstimated(usage: OcxUsage, usageStatus: UsageStatus, priceStatus: Exp
   return usage.estimated === true || usageStatus === "estimated" || priceStatus === "verified-derived";
 }
 
-/**
- * OpenAI provider ids eligible for service_tier "priority" price multipliers.
- * Only canonical OpenAI forward providers use the priority tier; routed providers
- * (OpenRouter, Cursor, etc.) may share model slugs but have independent pricing.
- */
-const OPENAI_TIER_PROVIDER_IDS = new Set(["openai", "openai-apikey"]);
-
 /**
  * Resolve the effective service tier from persisted log fields.
  * Priority: responseServiceTier (server-confirmed) > requestedServiceTier
@@ -405,9 +413,9 @@ function isConfirmedFast(tier?: ServiceTierInput): boolean {
  * normalized billable input — normalization subtracts cache read/write, so a
  * cache-heavy long prompt would fall below the boundary and under-bill.
  *
- * Skipped entirely for a response-confirmed Fast request: OpenAI does not serve
- * long context in Fast mode, so the two are mutually exclusive regimes rather
- * than composable multipliers.
+ * A provider's declaration decides how a response-confirmed priority tier relates to this band.
+ * OpenAI declares the bands exclusive. xAI publishes neither a combined rate nor an exclusion,
+ * so its long-context rate remains the known lower bound instead of inventing a stacked multiplier.
  */
 function applyContextTier(
   cost4: Cost4,
@@ -415,25 +423,27 @@ function applyContextTier(
   modelId: string,
   rawInputTokens: number | undefined,
   tier?: ServiceTierInput,
-): [Cost4, ContextTierName | undefined] {
-  if (rawInputTokens === undefined) return [cost4, undefined];
-  if (isConfirmedFast(tier)) return [cost4, undefined];
+): [Cost4, ContextTierName | undefined, boolean] {
+  if (rawInputTokens === undefined) return [cost4, undefined, false];
   const rule = findContextTier(baseProviderLabel(provider), modelId);
-  if (!rule || !isLongContext(rule, rawInputTokens)) return [cost4, undefined];
+  if (!rule || !isLongContext(rule, rawInputTokens)) return [cost4, undefined, false];
+  const confirmedFast = isConfirmedFast(tier);
+  if (confirmedFast && rule.confirmedPriorityRelation === "exclusive") {
+    return [cost4, undefined, false];
+  }
   return [{
     input: cost4.input * rule.multiplier.input,
     output: cost4.output * rule.multiplier.output,
     cacheRead: cost4.cacheRead * rule.multiplier.cacheRead,
     cacheWrite: cost4.cacheWrite * rule.multiplier.cacheWrite,
-  }, "long"];
+  }, "long", confirmedFast && rule.confirmedPriorityRelation === "lower-bound"];
 }
 
 /**
- * Apply the OpenAI priority-tier multiplier to a Cost4 when applicable.
+ * Apply a declared provider/model priority-tier multiplier to a Cost4 when applicable.
  * Returns [effectiveCost4, multiplier]. Multiplier is 1 (no-op) when:
  * - serviceTier is not "priority"
- * - provider is not a canonical OpenAI forward provider
- * - model is not in PRIORITY_MULTIPLIERS
+ * - no exact provider/model rule exists
  */
 function applyPriorityMultiplier(
   cost4: Cost4,
@@ -443,8 +453,9 @@ function applyPriorityMultiplier(
 ): [Cost4, number] {
   if (tierScalar(serviceTier) !== "priority") return [cost4, 1];
   const base = baseProviderLabel(provider);
-  if (!OPENAI_TIER_PROVIDER_IDS.has(base)) return [cost4, 1];
-  const multiplier = resolvePriorityMultiplier(modelId);
+  const rule = findPriorityPricingRule(base, modelId);
+  if (rule?.requiresResponseConfirmation && !isConfirmedFast(serviceTier)) return [cost4, 1];
+  const multiplier = rule?.multiplier ?? 1;
   if (multiplier === 1) return [cost4, 1];
   return [{
     input: cost4.input * multiplier,
@@ -495,16 +506,17 @@ export function estimateAttemptCost(
   const attemptServiceTier = attempt.tierOutcome
     ? serviceTierContextFromOutcome(attempt.tierOutcome)
     : serviceTier;
-  const [tieredCost4, contextTier] = applyContextTier(
+  const [tieredCost4, contextTier, contextPriorityLowerBound] = applyContextTier(
     price.cost4, attempt.provider, attempt.model, attempt.usage.inputTokens, attemptServiceTier,
   );
-  // Exclusive both ways: if the long rate applied, the request was NOT served as
-  // Fast (Fast does not support long context), so the Fast multiplier must not
-  // also apply — otherwise a downgraded request bills at both rates.
+  // A published long-context row owns the numeric estimate. OpenAI declares that band
+  // exclusive with Fast; xAI's confirmed combination is deliberately left unmultiplied
+  // and marked as a lower bound because no combined price has been published.
   const [effectiveCost4, multiplier] = contextTier
     ? [tieredCost4, 1] as const
     : applyPriorityMultiplier(tieredCost4, attempt.provider, attempt.model, attemptServiceTier);
-  const priorityLowerBound = isOpenRouterPriorityLowerBound(attempt.provider, attempt.tierOutcome);
+  const priorityLowerBound = contextPriorityLowerBound
+    || isOpenRouterPriorityLowerBound(attempt.provider, attempt.tierOutcome);
   return {
     ordinal: attempt.ordinal,
     provider: attempt.provider,
@@ -514,8 +526,8 @@ export function estimateAttemptCost(
     cost: calculateCost(tokens, effectiveCost4),
     estimated: isEstimated(attempt.usage, attempt.usageStatus, price.status),
     ...(multiplier !== 1 ? { priorityMultiplier: multiplier } : {}),
-    ...(priorityLowerBound ? { priorityLowerBound: true } : {}),
     ...(contextTier ? { contextTier } : {}),
+    ...(priorityLowerBound ? { priorityLowerBound: true } : {}),
   };
 }
 
@@ -557,8 +569,10 @@ export function estimateComboCost(
     ...(estimates.some(est => est.priorityMultiplier && est.priorityMultiplier !== 1)
       ? { priorityMultiplier: estimates.find(est => est.priorityMultiplier)?.priorityMultiplier }
       : {}),
-    ...(estimates.some(est => est.priorityLowerBound) ? { priorityLowerBound: true } : {}),
     ...(estimates.some(est => est.contextTier) ? { contextTier: "long" as const } : {}),
+    ...(estimates.every(est => est.priorityLowerBound === true)
+      ? { priorityLowerBound: true as const }
+      : {}),
   };
 }
 
@@ -579,13 +593,13 @@ export function estimateRequestCost(
   if (!tokens) return null;
   const price = resolveMatchedPrice(input.provider, input.model, overlays, userOverlays);
   if (!price) return null;
-  const [tieredCost4, contextTier] = applyContextTier(
+  const [tieredCost4, contextTier, contextPriorityLowerBound] = applyContextTier(
     price.cost4, input.provider, input.model, input.usage.inputTokens, input.serviceTier,
   );
   const [effectiveCost4, multiplier] = contextTier
     ? [tieredCost4, 1] as const
     : applyPriorityMultiplier(tieredCost4, input.provider, input.model, input.serviceTier);
-  const priorityLowerBound = isOpenRouterPriorityLowerBound(
+  const priorityLowerBound = contextPriorityLowerBound || isOpenRouterPriorityLowerBound(
     input.provider,
     typeof input.serviceTier === "object" ? input.serviceTier.tierOutcome : undefined,
   );
@@ -595,8 +609,8 @@ export function estimateRequestCost(
     cost: calculateCost(tokens, effectiveCost4),
     estimated: isEstimated(input.usage, input.usageStatus, price.status),
     ...(multiplier !== 1 ? { priorityMultiplier: multiplier } : {}),
-    ...(priorityLowerBound ? { priorityLowerBound: true } : {}),
     ...(contextTier ? { contextTier } : {}),
+    ...(priorityLowerBound ? { priorityLowerBound: true } : {}),
   };
 }
 
diff --git a/src/usage/expected-prices.ts b/src/usage/expected-prices.ts
index 64a42e87e7..af94b8b262 100644
--- a/src/usage/expected-prices.ts
+++ b/src/usage/expected-prices.ts
@@ -181,6 +181,30 @@ export const EXPECTED_PRICE_OVERLAYS: readonly ExpectedPriceOverlay[] = [
   { provider: "cursor", modelId: "auto", cost4: { input: 1.25, output: 6, cacheRead: 0.25, cacheWrite: 1.25 }, source: "https://docs.cursor.com/account/pricing + https://cursor.com/blog/aug-2025-pricing", verifiedAt: "2026-07-20", status: "verified" },
 ];
 
+/**
+ * Exact official corrections for stale nonzero catalog rows. These are intentionally separate
+ * from fallback overlays: they win over the bundled row only for the declared provider/model and
+ * therefore cannot reprice routed resellers that reuse the same model slug.
+ */
+export const VERIFIED_PRICE_OVERRIDES: readonly ExpectedPriceOverlay[] = [
+  {
+    provider: "xai",
+    modelId: "grok-4.6",
+    cost4: { input: 2, output: 6, cacheRead: 0.5, cacheWrite: 0 },
+    source: "https://docs.x.ai/developers/pricing",
+    verifiedAt: "2026-08-18",
+    status: "verified",
+  },
+];
+
+export function findVerifiedPriceOverride(
+  provider: string,
+  modelId: string,
+  overrides: readonly ExpectedPriceOverlay[] = VERIFIED_PRICE_OVERRIDES,
+): ExpectedPriceOverlay | undefined {
+  return overrides.find(row => row.provider === provider && row.modelId === modelId);
+}
+
 /**
  * Exact-key overlay lookup. Returns verified first, then verified-derived.
  * NEVER returns "unverified" rows — fail-closed is enforced in code, not just docs.
@@ -196,12 +220,7 @@ export function findExpectedPriceOverlay(
     ?? exact.find(row => row.status === "verified-derived");
 }
 
-/**
- * OpenAI Fast mode (`service_tier=priority`) price multipliers by model slug.
- * Source: https://openai.com/api-fast-mode/ (2026-07-31).
- * Fast pricing applies uniformly to all token types (input, output, cache).
- * Models not listed here fall back to 1× (no multiplier).
- */
+/** OpenAI Fast price multipliers retained as a compatibility export. */
 export const PRIORITY_MULTIPLIERS: Readonly<Record<string, number>> = {
   "gpt-5.6-sol": 2,
   // Post-price-cut Fast tables (https://openai.com/api-fast-mode/, 2026-08-05):
@@ -219,6 +238,52 @@ export function resolvePriorityMultiplier(modelId: string): number {
   return PRIORITY_MULTIPLIERS[modelId] ?? 1;
 }
 
+export interface PriorityPricingRule {
+  provider: string;
+  modelId: string;
+  multiplier: number;
+  /** Apply the premium only after the upstream response confirms this tier. */
+  requiresResponseConfirmation?: true;
+  source: string;
+  verifiedAt: string;
+}
+
+const OPENAI_FAST_PRICING = "https://openai.com/api-fast-mode/";
+const XAI_PRIORITY_PRICING = "https://docs.x.ai/developers/advanced-api-usage/priority-processing";
+
+/**
+ * Exact provider/model priority premiums. Routed resellers never inherit a vendor rule merely
+ * because they reuse its model slug. Multipliers apply uniformly after cache discounts.
+ */
+export const PRIORITY_PRICING_RULES: readonly PriorityPricingRule[] = [
+  ...["openai", "openai-apikey"].flatMap(provider =>
+    Object.entries(PRIORITY_MULTIPLIERS).map(([modelId, multiplier]): PriorityPricingRule => ({
+      provider,
+      modelId,
+      multiplier,
+      source: OPENAI_FAST_PRICING,
+      verifiedAt: "2026-08-05",
+    })),
+  ),
+  ...["grok-4.5", "grok-4.6"].map((modelId): PriorityPricingRule => ({
+    provider: "xai",
+    modelId,
+    multiplier: 2,
+    requiresResponseConfirmation: true,
+    source: XAI_PRIORITY_PRICING,
+    verifiedAt: "2026-08-18",
+  })),
+];
+
+/** Exact provider/model priority-pricing lookup. */
+export function findPriorityPricingRule(
+  provider: string,
+  modelId: string,
+  rules: readonly PriorityPricingRule[] = PRIORITY_PRICING_RULES,
+): PriorityPricingRule | undefined {
+  return rules.find(rule => rule.provider === provider && rule.modelId === modelId);
+}
+
 /**
  * Long-context pricing tiers (#908). Several vendors reprice the ENTIRE request
  * once the prompt crosses a published input-token threshold, so a flat Cost4
@@ -244,6 +309,8 @@ export interface ContextTier {
   inclusive: boolean;
   /** Per-field factor from the short rate to the published long rate. */
   multiplier: Cost4;
+  /** Published relationship between confirmed priority and long-context bands. */
+  confirmedPriorityRelation?: "exclusive" | "lower-bound";
   source: string;
   verifiedAt: string;
 }
@@ -277,6 +344,7 @@ export const CONTEXT_TIERS: readonly ContextTier[] = [
       thresholdInputTokens: 272_000,
       inclusive: false,
       multiplier: OPENAI_LONG_CONTEXT,
+      confirmedPriorityRelation: "exclusive",
       source: OPENAI_PRICING_DOC,
       verifiedAt: "2026-08-03",
     })),
@@ -287,19 +355,21 @@ export const CONTEXT_TIERS: readonly ContextTier[] = [
     thresholdInputTokens: 200_000,
     inclusive: true,
     multiplier: UNIFORM_DOUBLE,
+    confirmedPriorityRelation: "lower-bound",
     source: "https://docs.x.ai/developers/pricing",
     verifiedAt: "2026-08-03",
   },
   {
-    // 260813: grok-4.6 long-context tier mirrored from grok-4.5; the official pricing row
-    // was not yet published when the model page went up, so treat as provisional.
+    // xAI publishes the whole-request >=200k band for grok-4.6. Its combination with
+    // Priority Processing is not published, so confirmed priority uses this row as a lower bound.
     provider: "xai",
     modelId: "grok-4.6",
     thresholdInputTokens: 200_000,
     inclusive: true,
     multiplier: UNIFORM_DOUBLE,
+    confirmedPriorityRelation: "lower-bound",
     source: "https://docs.x.ai/developers/pricing",
-    verifiedAt: "2026-08-13",
+    verifiedAt: "2026-08-18",
   },
   {
     // daybreak-blue-latest aliases gpt-5.6-sol, which publishes the full long-context row
diff --git a/tests/fastwire-policy.test.ts b/tests/fastwire-policy.test.ts
index 412d0524a4..a29debcb34 100644
--- a/tests/fastwire-policy.test.ts
+++ b/tests/fastwire-policy.test.ts
@@ -288,6 +288,7 @@ describe("resolveFastPolicy matrix", () => {
         settledCallerTier: undefined,
       },
       {
+        // B2: key-auth Chat Completions is a documented Priority Processing transport.
         name: "xAI API-key default",
         providerName: "xai",
         modelIds: ["grok-4.6", "grok-4.5"],
@@ -297,7 +298,7 @@ describe("resolveFastPolicy matrix", () => {
           authMode: "key" as const,
         },
         adapter: "openai-chat",
-        forwardCallerTier: false,
+        forwardCallerTier: true,
         callerTier: undefined,
         settledCallerTier: undefined,
       },
diff --git a/tests/management-api-logs-metrics.test.ts b/tests/management-api-logs-metrics.test.ts
index f7077a8951..0255005e2f 100644
--- a/tests/management-api-logs-metrics.test.ts
+++ b/tests/management-api-logs-metrics.test.ts
@@ -95,6 +95,31 @@ describe("GET /api/logs display metrics", () => {
     expect(dto!.displayMetrics.cost.estimateReasons).toContain("cache_detail_missing");
   });
 
+  test("confirmed xAI priority plus long context is exposed as a cost lower bound", async () => {
+    addRequestLog(baseEntry({
+      provider: "xai",
+      model: "grok-4.6",
+      usage: {
+        inputTokens: 200_000,
+        outputTokens: 10_000,
+        cacheReadInputTokens: 50_000,
+      },
+      tierOutcome: {
+        canonical: "priority",
+        wireKind: "service-tier",
+        wireValue: "priority",
+        fastOutcome: "applied",
+        confirmation: "confirmed",
+        responseServiceTier: "priority",
+      },
+    }));
+    const [dto] = await readLogs();
+    expect(dto!.displayMetrics.cost.kind).toBe("value");
+    expect(dto!.displayMetrics.cost.estimate.priorityLowerBound).toBe(true);
+    expect(dto!.displayMetrics.cost.estimate.cost.total).toBeCloseTo(0.77, 9);
+    expect(dto!.displayMetrics.cost.estimateReasons).toContain("priority_lower_bound");
+  });
+
   test("unmatched price is unavailable instead of zero", async () => {
     addRequestLog(baseEntry({
       provider: "no-such-provider",
diff --git a/tests/service-tier-capability.test.ts b/tests/service-tier-capability.test.ts
index 49c051b8dc..fd1a418602 100644
--- a/tests/service-tier-capability.test.ts
+++ b/tests/service-tier-capability.test.ts
@@ -7,15 +7,22 @@
  * ever receiving an injection (PR #860 family).
  */
 import { afterEach, describe, expect, test } from "bun:test";
-import { applyProviderConfigHints } from "../src/codex/catalog";
+import { applyProviderConfigHints, buildCatalogEntries, gatherRoutedModels } from "../src/codex/catalog";
 import { applyCatalogModelMetadata } from "../src/codex/catalog/effort";
 import type { RawEntry } from "../src/codex/catalog/parsing";
 import { providerConfigSeed, enrichProviderFromRegistry } from "../src/providers/derive";
 import { getProviderRegistryEntry } from "../src/providers/registry";
+import { decideTier } from "../src/providers/fastwire";
 import type { RequestLogContext } from "../src/server/request-log";
 import { applyServiceTierGate, handleResponses } from "../src/server/responses/core";
-import { canForwardServiceTierForModel, serviceTierSupportForModel, supportsServiceTierForModel } from "../src/providers/service-tier";
-import { serviceTierAdapterForModel } from "../src/providers/service-tier";
+import {
+  canForwardServiceTierForModel,
+  fastPolicyForModel,
+  serviceTierAdapterForModel,
+  serviceTierSupportForModel,
+  serviceTierSupportFromPolicy,
+  supportsServiceTierForModel,
+} from "../src/providers/service-tier";
 import { candidateCapabilityEvidence } from "../src/routing/capability";
 import { resolveProductionBehaviorValues } from "../src/routing/compatibility/behavior";
 import type { OcxConfig, OcxProviderConfig } from "../src/types";
@@ -74,6 +81,92 @@ describe("registry capability reaches saved configs without overriding them", ()
   });
 });
 
+describe("xAI Fast capability follows the captured authentication transport", () => {
+  function xaiProvider(
+    authMode: "key" | "oauth",
+    overrides: Partial<OcxProviderConfig> = {},
+  ): OcxProviderConfig {
+    return {
+      ...providerConfigSeed(getProviderRegistryEntry("xai")!),
+      authMode,
+      apiKey: authMode === "key" ? "xai-test-key" : "oauth-test-token",
+      liveModels: false,
+      models: ["grok-4.6"],
+      ...overrides,
+    };
+  }
+
+  async function catalogEntry(provider: OcxProviderConfig) {
+    const models = await gatherRoutedModels({
+      providers: { xai: provider },
+    } as unknown as OcxConfig);
+    return buildCatalogEntries(null, [], models)
+      .find(entry => entry.slug === "xai/grok-4.6");
+  }
+
+  test("registry declares a key-auth overlay without classifying OAuth", () => {
+    const entry = getProviderRegistryEntry("xai")!;
+    expect(entry.keyAuthServiceTier).toEqual({
+      supportsServiceTier: true,
+      chatServiceTier: true,
+    });
+    expect(entry.supportsServiceTier).toBeUndefined();
+    expect(entry.chatServiceTier).toBeUndefined();
+
+    const keyPolicy = fastPolicyForModel(xaiProvider("key"), "grok-4.6", "xai");
+    expect(keyPolicy).toMatchObject({
+      capability: true,
+      eligibility: "eligible",
+      forwardCallerTier: true,
+      fastTierDescription: "Priority processing, 2x token price",
+    });
+
+    const oauthPolicy = fastPolicyForModel(xaiProvider("oauth"), "grok-4.6", "xai");
+    expect(oauthPolicy.capability).toBeUndefined();
+    expect(oauthPolicy.eligibility).toBe("unclassified");
+    expect(oauthPolicy.forwardCallerTier).toBe(false);
+  });
+
+  test("catalog and runtime publish the same key/OAuth conclusion", async () => {
+    const keyProvider = xaiProvider("key");
+    const keyPolicy = fastPolicyForModel(keyProvider, "grok-4.6", "xai");
+    const keyCatalog = await catalogEntry(keyProvider);
+    expect(serviceTierSupportFromPolicy(keyPolicy)).toBe(true);
+    expect(keyCatalog?.service_tiers).toEqual([{
+      id: "priority",
+      name: "Fast",
+      description: "Priority processing, 2x token price",
+    }]);
+    expect(keyCatalog?.additional_speed_tiers).toEqual(["fast"]);
+    expect(decideTier(keyPolicy, true, undefined)).toEqual({ kind: "set", value: "priority" });
+
+    const oauthProvider = xaiProvider("oauth");
+    const oauthPolicy = fastPolicyForModel(oauthProvider, "grok-4.6", "xai");
+    const oauthCatalog = await catalogEntry(oauthProvider);
+    expect(serviceTierSupportFromPolicy(oauthPolicy)).toBe(false);
+    expect(oauthCatalog).not.toHaveProperty("service_tiers");
+    expect(oauthCatalog).not.toHaveProperty("additional_speed_tiers");
+    expect(decideTier(oauthPolicy, true, undefined)).toEqual({ kind: "drop" });
+  });
+
+  test("explicit supportsServiceTier=false wins in policy and catalog for both transports", async () => {
+    for (const authMode of ["key", "oauth"] as const) {
+      const provider = xaiProvider(authMode, { supportsServiceTier: false });
+      const policy = fastPolicyForModel(
+        provider,
+        "grok-4.6",
+        "xai",
+      );
+      expect(policy.capability).toBe(false);
+      expect(policy.eligibility).toBe("capability-unsupported");
+      expect(decideTier(policy, true, undefined)).toEqual({ kind: "drop" });
+      const catalog = await catalogEntry(provider);
+      expect(catalog).not.toHaveProperty("service_tiers");
+      expect(catalog).not.toHaveProperty("additional_speed_tiers");
+    }
+  });
+});
+
 describe("service-tier capability is exact-model and provider-scoped", () => {
   test("an exact model entry overrides the provider fallback in both directions", () => {
     const provider: OcxProviderConfig = {
@@ -264,6 +357,16 @@ describe("the gate fires on the live handleResponses path", () => {
     ({ ...providerConfigSeed(getProviderRegistryEntry("deepseek")!), apiKey: "sk-test" });
   const openAiKeyProvider = (): OcxProviderConfig =>
     ({ ...providerConfigSeed(getProviderRegistryEntry("openai-apikey")!), apiKey: "sk-test" });
+  const xaiKeyProvider = (): OcxProviderConfig => ({
+    ...providerConfigSeed(getProviderRegistryEntry("xai")!),
+    authMode: "key",
+    apiKey: "xai-test-key",
+  });
+  const xaiOAuthProvider = (): OcxProviderConfig => ({
+    ...providerConfigSeed(getProviderRegistryEntry("xai")!),
+    authMode: "oauth",
+    apiKey: "xai-oauth-test-token",
+  });
   const openRouterProvider = (overrides: Partial<OcxProviderConfig> = {}): OcxProviderConfig => {
     const provider: OcxProviderConfig = {
       ...providerConfigSeed(getProviderRegistryEntry("openrouter")!),
@@ -321,6 +424,23 @@ describe("the gate fires on the live handleResponses path", () => {
     expect(body.service_tier).toBe("flex");
   });
 
+  test("xAI API-key runtime injects priority while OAuth does not", async () => {
+    const keyBody = await drive("xai", xaiKeyProvider(), "grok-4.6", {}, true);
+    expect(keyBody.service_tier).toBe("priority");
+    const oauthBody = await drive("xai", xaiOAuthProvider(), "grok-4.6", {}, true);
+    expect(oauthBody).not.toHaveProperty("service_tier");
+    for (const provider of [xaiKeyProvider(), xaiOAuthProvider()]) {
+      const optedOut = await drive(
+        "xai",
+        { ...provider, supportsServiceTier: false },
+        "grok-4.6",
+        {},
+        true,
+      );
+      expect(optedOut).not.toHaveProperty("service_tier");
+    }
+  });
+
   test("an unclassified custom Responses provider keeps caller values; only explicit false strips", async () => {
     const custom = (): OcxProviderConfig => ({ adapter: "openai-responses", baseUrl: "https://gateway.example.com/v1", apiKey: "sk-test" });
     const preserved = await drive("custom-gw", custom(), "some-model", { service_tier: "priority" });
diff --git a/tests/usage-cost.test.ts b/tests/usage-cost.test.ts
index 11fa3d77fc..362868f4f2 100644
--- a/tests/usage-cost.test.ts
+++ b/tests/usage-cost.test.ts
@@ -1,4 +1,5 @@
 import { afterEach, describe, expect, test } from "bun:test";
+import { createAdapterTierMetadata } from "../src/providers/fastwire";
 import {
   calculateCost,
   estimateAttemptCost,
@@ -12,8 +13,10 @@ import {
 import {
   EXPECTED_PRICE_OVERLAYS,
   PRIORITY_MULTIPLIERS,
+  PRIORITY_PRICING_RULES,
   CONTEXT_TIERS,
   findExpectedPriceOverlay,
+  findPriorityPricingRule,
   resolvePriorityMultiplier,
   type ExpectedPriceOverlay,
 } from "../src/usage/expected-prices";
@@ -564,6 +567,169 @@ describe("priority (Fast) service tier multiplier", () => {
   });
 });
 
+describe("xAI Priority Processing pricing", () => {
+  const usage = {
+    inputTokens: 100_000,
+    outputTokens: 10_000,
+    cacheReadInputTokens: 20_000,
+  };
+
+  function outcome(responseServiceTier?: string) {
+    const tracker = createAdapterTierMetadata(
+      {
+        capability: true,
+        eligibility: "eligible",
+        fastWire: {
+          kind: "service-tier",
+          canonicalToWire: { priority: "priority" },
+          foreignCallerTiers: "verbatim",
+        },
+        demandDecision: "force-fast",
+      },
+      { kind: "set", value: "priority" },
+      "service-tier",
+      "priority",
+    )!;
+    if (responseServiceTier !== undefined) tracker.observeResponseServiceTier(responseServiceTier);
+    return tracker.outcome;
+  }
+
+  function estimate(tierOutcome: ReturnType<typeof outcome>, requestUsage = usage) {
+    return estimateAttemptCost({
+      ordinal: 1,
+      provider: "xai",
+      model: "grok-4.6",
+      usageStatus: "reported",
+      usage: requestUsage,
+      tierOutcome,
+    })!;
+  }
+
+  test("xAI rules declare exact 2x premiums with official provenance", () => {
+    const xaiRules = PRIORITY_PRICING_RULES.filter(rule => rule.provider === "xai");
+    expect(xaiRules.map(rule => rule.modelId)).toEqual(["grok-4.5", "grok-4.6"]);
+    expect(xaiRules.every(rule => rule.multiplier === 2)).toBe(true);
+    expect(xaiRules.every(rule => rule.requiresResponseConfirmation === true)).toBe(true);
+    expect(xaiRules.every(rule => rule.source === "https://docs.x.ai/developers/advanced-api-usage/priority-processing")).toBe(true);
+    expect(findPriorityPricingRule("xai", "grok-4.6")?.multiplier).toBe(2);
+    expect(findPriorityPricingRule("openrouter", "grok-4.6")).toBeUndefined();
+    expect(resolveMatchedPrice("openrouter", "grok-4.6")?.cost4).toEqual({
+      input: 2,
+      output: 6,
+      cacheRead: 0.3,
+      cacheWrite: 0,
+    });
+    expect(resolveMatchedPrice("cursor", "grok-4.6")?.cost4).toEqual({
+      input: 2,
+      output: 6,
+      cacheRead: 0.3,
+      cacheWrite: 0,
+    });
+  });
+
+  test("grok-4.6 standard and confirmed priority prices include the official cache rate", () => {
+    expect(resolveMatchedPrice("xai", "grok-4.6")?.cost4).toEqual({
+      input: 2,
+      output: 6,
+      cacheRead: 0.5,
+      cacheWrite: 0,
+    });
+    const confirmedOutcome = outcome("priority");
+    const confirmed = estimate(confirmedOutcome);
+    expect(confirmedOutcome).toMatchObject({
+      canonical: "priority",
+      fastOutcome: "applied",
+      confirmation: "confirmed",
+    });
+    expect(confirmed.cost.total).toBeCloseTo(0.46, 9);
+    expect(confirmed.cost.cacheRead).toBeCloseTo(0.02, 9);
+    expect(confirmed.priorityMultiplier).toBe(2);
+  });
+
+  test("an assumed priority outcome stays at the standard price", () => {
+    const assumedOutcome = outcome();
+    const assumed = estimate(assumedOutcome);
+    expect(assumedOutcome).toMatchObject({
+      canonical: "priority",
+      fastOutcome: "applied",
+      confirmation: "assumed",
+    });
+    expect(assumed.cost.total).toBeCloseTo(0.23, 9);
+    expect(assumed.priorityMultiplier).toBeUndefined();
+  });
+
+  test("missing provenance and a requested tier do not prove the xAI premium", () => {
+    for (const serviceTier of [
+      "priority",
+      { requestedServiceTier: "priority" },
+      { configuredServiceTier: "priority" },
+    ] as const) {
+      const unconfirmed = estimateRequestCost({
+        provider: "xai",
+        model: "grok-4.6",
+        usageStatus: "reported",
+        usage,
+        serviceTier,
+      })!;
+      expect(unconfirmed.cost.total).toBeCloseTo(0.23, 9);
+      expect(unconfirmed.priorityMultiplier).toBeUndefined();
+    }
+  });
+
+  test("an echoed default records a downgrade and bills the standard price", () => {
+    const downgradedOutcome = outcome("default");
+    const downgraded = estimate(downgradedOutcome);
+    expect(downgradedOutcome).toMatchObject({
+      fastOutcome: "downgraded",
+      fastDowngradeReason: "response-declined",
+      confirmation: "downgraded",
+      responseServiceTier: "default",
+    });
+    expect(downgradedOutcome).not.toHaveProperty("canonical");
+    expect(downgraded.cost.total).toBeCloseTo(0.23, 9);
+    expect(downgraded.priorityMultiplier).toBeUndefined();
+  });
+
+  test("confirmed priority at 200k uses the long-context price as a marked lower bound", () => {
+    const long = estimate(outcome("priority"), {
+      inputTokens: 200_000,
+      outputTokens: 10_000,
+      cacheReadInputTokens: 50_000,
+    });
+    expect(long.contextTier).toBe("long");
+    expect(long.priorityMultiplier).toBeUndefined();
+    expect(long.priorityLowerBound).toBe(true);
+    expect(long.cost).toMatchObject({
+      input: 0.6,
+      cacheRead: 0.05,
+      output: 0.12,
+    });
+    expect(long.cost.total).toBeCloseTo(0.77, 9);
+  });
+
+  test("a combo is a lower bound only when every priced attempt is a lower bound", () => {
+    const confirmed = outcome("priority");
+    const lowerBoundAttempt = {
+      ordinal: 1,
+      provider: "xai",
+      model: "grok-4.6",
+      usageStatus: "reported" as const,
+      usage: { inputTokens: 200_000, outputTokens: 10_000 },
+      tierOutcome: confirmed,
+    };
+    const ordinaryAttempt = {
+      ordinal: 2,
+      provider: "xai",
+      model: "grok-4.6",
+      usageStatus: "reported" as const,
+      usage,
+    };
+
+    expect(estimateComboCost([lowerBoundAttempt, { ...lowerBoundAttempt, ordinal: 2 }])?.priorityLowerBound).toBe(true);
+    expect(estimateComboCost([lowerBoundAttempt, ordinaryAttempt])?.priorityLowerBound).toBeUndefined();
+  });
+});
+
 describe("long-context pricing tiers (#908)", () => {
   const SOL: ExpectedPriceOverlay[] = [
     { provider: "openai", modelId: "gpt-5.6-sol", cost4: { input: 5, output: 30, cacheRead: 0.5, cacheWrite: 6.25 }, source: "test", verifiedAt: "2026-08-03", status: "verified" },