diff --git a/src/app-server/service-tier.ts b/src/app-server/service-tier.ts
index e1380741..5991fb1a 100644
--- a/src/app-server/service-tier.ts
+++ b/src/app-server/service-tier.ts
@@ -2,13 +2,21 @@ export type ServiceTier = string;
 export type RequestedServiceTier = "fast" | "off";
 export type ServiceTierRequest = string | null | undefined;
 
+export interface ServiceTierMetadata {
+  id: string;
+  name: string;
+}
+
 export function parseServiceTier(value: unknown): ServiceTier | null {
   if (typeof value === "string" && value.length > 0) return value;
   return null;
 }
 
-export function requestedServiceTierRequestValue(value: RequestedServiceTier | null): "fast" | null | undefined {
-  if (value === "fast") return "fast";
+export function requestedServiceTierRequestValue(
+  value: RequestedServiceTier | null,
+  fastServiceTierId = "fast",
+): string | null | undefined {
+  if (value === "fast") return fastServiceTierId;
   if (value === "off") return null;
   return undefined;
 }
@@ -16,3 +24,10 @@ export function requestedServiceTierRequestValue(value: RequestedServiceTier | n
 export function configuredServiceTierRequestValue(value: ServiceTier | null): string | undefined {
   return value ?? undefined;
 }
+
+export function isFastServiceTier(value: ServiceTier | null | undefined, serviceTiers: readonly ServiceTierMetadata[] = []): boolean {
+  if (!value) return false;
+  if (value === "fast") return true;
+  if (serviceTiers.length === 0) return value === "priority";
+  return serviceTiers.some((tier) => tier.id === value && tier.name.trim().toLowerCase() === "fast");
+}
diff --git a/src/features/chat/runtime-settings-controller.ts b/src/features/chat/runtime-settings-controller.ts
index 47e2d00c..37aa9d3d 100644
--- a/src/features/chat/runtime-settings-controller.ts
+++ b/src/features/chat/runtime-settings-controller.ts
@@ -8,7 +8,8 @@ import { collaborationModeToggleMessage, nextCollaborationMode } from "../../run
 import { readRuntimeConfig } from "../../runtime/config";
 import {
   autoReviewActive,
-  currentServiceTier,
+  fastModeActive,
+  fastServiceTierRequestValue,
   requestedTurnRuntimeSettings,
   runtimeOverridePayload,
   type RuntimeSnapshot,
@@ -78,8 +79,9 @@ export class ChatRuntimeSettingsController {
   }
 
   async toggleFastMode(): Promise<void> {
-    const current = currentServiceTier(this.host.runtimeSnapshot(), readRuntimeConfig(this.state.effectiveConfig));
-    const next: RequestedServiceTier = current === "fast" ? "off" : "fast";
+    const snapshot = this.host.runtimeSnapshot();
+    const config = readRuntimeConfig(this.state.effectiveConfig);
+    const next: RequestedServiceTier = fastModeActive(snapshot, config) ? "off" : "fast";
     this.dispatch({ type: "runtime/requested-service-tier-set", serviceTier: next });
     this.dispatch({ type: "ui/panel-set", panel: null });
     if (!(await this.applyPendingThreadSettings())) return;
@@ -112,7 +114,8 @@ export class ChatRuntimeSettingsController {
   private pendingThreadSettingsUpdate(): ThreadSettingsUpdate {
     const update: ThreadSettingsUpdate = {};
     const state = this.state;
-    const turnSettings = requestedTurnRuntimeSettings(this.host.runtimeSnapshot());
+    const snapshot = this.host.runtimeSnapshot();
+    const turnSettings = requestedTurnRuntimeSettings(snapshot);
 
     if (state.requestedModel.kind !== "default") {
       const model = runtimeOverridePayload(state.requestedModel);
@@ -123,7 +126,10 @@ export class ChatRuntimeSettingsController {
       if (effort !== undefined) update.effort = effort;
     }
     if (state.requestedServiceTier !== null) {
-      const serviceTier = requestedServiceTierRequestValue(state.requestedServiceTier);
+      const serviceTier = requestedServiceTierRequestValue(
+        state.requestedServiceTier,
+        fastServiceTierRequestValue(snapshot, readRuntimeConfig(state.effectiveConfig)),
+      );
       if (serviceTier !== undefined) update.serviceTier = serviceTier;
     }
     if (state.requestedApprovalsReviewer !== null) {
diff --git a/src/features/chat/view-model.ts b/src/features/chat/view-model.ts
index a74f5e6c..57e6c183 100644
--- a/src/features/chat/view-model.ts
+++ b/src/features/chat/view-model.ts
@@ -5,7 +5,7 @@ import {
   autoReviewActive,
   currentModel,
   currentReasoningEffort,
-  currentServiceTier,
+  fastModeActive,
   runtimeOverrideLabel,
   runtimeSummaryLabel,
   serviceTierLabel,
@@ -156,7 +156,7 @@ export function toolbarViewModel(input: ToolbarViewModelInput): ToolbarViewModel
     runtimeOpen,
     planActive: state.requestedCollaborationMode === "plan",
     autoReviewActive: autoReviewActive(snapshot, config),
-    fastActive: currentServiceTier(snapshot, config) === "fast",
+    fastActive: fastModeActive(snapshot, config),
     runtimeSummary: runtimeSummaryLabel(model, effort),
     runtimeTitle: `Model: ${model ?? "(Codex default)"}; Effort: ${effort ?? "(Codex default)"}`,
     runtimeEmphasized: false,
diff --git a/src/runtime/state.ts b/src/runtime/state.ts
index f2b074ea..9dad0594 100644
--- a/src/runtime/state.ts
+++ b/src/runtime/state.ts
@@ -10,6 +10,7 @@ import type { RateLimitSnapshot } from "../generated/app-server/v2/RateLimitSnap
 import type { ThreadTokenUsage } from "../generated/app-server/v2/ThreadTokenUsage";
 import {
   configuredServiceTierRequestValue,
+  isFastServiceTier,
   requestedServiceTierRequestValue,
   type RequestedServiceTier,
   type ServiceTier,
@@ -61,7 +62,9 @@ export function requestedOrConfiguredServiceTier(
   snapshot: RuntimeSnapshot,
   config: RuntimeConfigProjection = readRuntimeConfig(snapshot.effectiveConfig),
 ): ServiceTierRequest {
-  if (snapshot.requestedServiceTier !== null) return requestedServiceTierRequestValue(snapshot.requestedServiceTier);
+  if (snapshot.requestedServiceTier !== null) {
+    return requestedServiceTierRequestValue(snapshot.requestedServiceTier, fastServiceTierRequestValue(snapshot, config));
+  }
   return configuredServiceTierRequestValue(config.serviceTier);
 }
 
@@ -138,16 +141,30 @@ export function serviceTierLabel(
   return currentServiceTier(snapshot, config) ?? "(Codex default)";
 }
 
+export function fastModeActive(
+  snapshot: RuntimeSnapshot,
+  config: RuntimeConfigProjection = readRuntimeConfig(snapshot.effectiveConfig),
+): boolean {
+  return isFastServiceTier(currentServiceTier(snapshot, config), currentModelServiceTiers(snapshot, config));
+}
+
 export function fastModeLabel(
   snapshot: RuntimeSnapshot,
   config: RuntimeConfigProjection = readRuntimeConfig(snapshot.effectiveConfig),
 ): string {
   if (snapshot.requestedServiceTier === "off") return "off";
+  if (fastModeActive(snapshot, config)) return "on";
   const serviceTier = currentServiceTier(snapshot, config);
-  if (serviceTier === "fast") return "on";
   return serviceTier ? "off" : "Codex default";
 }
 
+export function fastServiceTierRequestValue(
+  snapshot: RuntimeSnapshot,
+  config: RuntimeConfigProjection = readRuntimeConfig(snapshot.effectiveConfig),
+): string {
+  return currentModelServiceTiers(snapshot, config).find((tier) => tier.name.trim().toLowerCase() === "fast")?.id ?? "fast";
+}
+
 export function defaultRuntimeOverride<T>(): RuntimeOverride<T> {
   return { kind: "default" };
 }
@@ -175,3 +192,10 @@ export function runtimeOverrideLabel<T>(override: RuntimeOverride<T>): string {
 function isAutoReviewReviewer(value: ApprovalsReviewer | null): boolean {
   return value === "auto_review" || value === "guardian_subagent";
 }
+
+function currentModelServiceTiers(
+  snapshot: RuntimeSnapshot,
+  config: RuntimeConfigProjection = readRuntimeConfig(snapshot.effectiveConfig),
+): Model["serviceTiers"] {
+  return findModelByIdOrName(snapshot.availableModels, currentModel(snapshot, config))?.serviceTiers ?? [];
+}
diff --git a/tests/app-server/service-tier.test.ts b/tests/app-server/service-tier.test.ts
index f1e723c4..e51ebac5 100644
--- a/tests/app-server/service-tier.test.ts
+++ b/tests/app-server/service-tier.test.ts
@@ -1,6 +1,11 @@
 import { describe, expect, it } from "vitest";
 
-import { configuredServiceTierRequestValue, parseServiceTier, requestedServiceTierRequestValue } from "../../src/app-server/service-tier";
+import {
+  configuredServiceTierRequestValue,
+  isFastServiceTier,
+  parseServiceTier,
+  requestedServiceTierRequestValue,
+} from "../../src/app-server/service-tier";
 
 describe("service tier", () => {
   it("accepts non-empty service tier ids from config and app-server reports", () => {
@@ -20,6 +25,7 @@ describe("service tier", () => {
 
   it("serializes panel fast mode off as an explicit null request", () => {
     expect(requestedServiceTierRequestValue("fast")).toBe("fast");
+    expect(requestedServiceTierRequestValue("fast", "priority")).toBe("priority");
     expect(requestedServiceTierRequestValue("off")).toBeNull();
     expect(requestedServiceTierRequestValue(null)).toBeUndefined();
   });
@@ -29,4 +35,23 @@ describe("service tier", () => {
     expect(configuredServiceTierRequestValue("catalog-tier")).toBe("catalog-tier");
     expect(configuredServiceTierRequestValue(null)).toBeUndefined();
   });
+
+  it("recognizes Codex fast tier aliases without rejecting other tier ids", () => {
+    expect(isFastServiceTier("fast")).toBe(true);
+    expect(isFastServiceTier("priority")).toBe(true);
+    expect(isFastServiceTier("catalog-fast", [{ id: "catalog-fast", name: "Fast" }])).toBe(true);
+    expect(isFastServiceTier("priority", [{ id: "priority", name: "Priority" }])).toBe(false);
+    expect(isFastServiceTier("flex", [{ id: "flex", name: "Flex" }])).toBe(false);
+  });
+
+  it("locks the Codex 0.134.0 Fast catalog semantics observed from app-server", () => {
+    // Codex app-server 0.134.0 reports Fast as serviceTiers[{ id: "priority", name: "Fast" }].
+    const codex01340FastTier = { id: "priority", name: "Fast" };
+
+    expect(requestedServiceTierRequestValue("fast", codex01340FastTier.id)).toBe("priority");
+    expect(isFastServiceTier("priority", [codex01340FastTier])).toBe(true);
+
+    // Clearing Fast with serviceTier: null is reported back by 0.134.0 as "default", not null.
+    expect(isFastServiceTier("default", [codex01340FastTier])).toBe(false);
+  });
 });
diff --git a/tests/features/chat/runtime-settings-controller.test.ts b/tests/features/chat/runtime-settings-controller.test.ts
index 83dbc21c..9442e67c 100644
--- a/tests/features/chat/runtime-settings-controller.test.ts
+++ b/tests/features/chat/runtime-settings-controller.test.ts
@@ -1,8 +1,9 @@
 import { describe, expect, it, vi } from "vitest";
 
 import { ChatRuntimeSettingsController } from "../../../src/features/chat/runtime-settings-controller";
-import { createChatState, createChatStateStore } from "../../../src/features/chat/chat-state";
+import { createChatState, createChatStateStore, type ChatAction } from "../../../src/features/chat/chat-state";
 import type { AppServerClient } from "../../../src/app-server/client";
+import type { Model } from "../../../src/generated/app-server/v2/Model";
 
 describe("ChatRuntimeSettingsController", () => {
   it("applies pending runtime overrides through thread settings and commits them", async () => {
@@ -62,6 +63,30 @@ describe("ChatRuntimeSettingsController", () => {
     expect(messages).toEqual(["Fast mode on for subsequent turns."]);
   });
 
+  it("requests the catalog Fast tier id and toggles it off from the reported effective id", async () => {
+    const state = createChatState();
+    state.activeThreadId = "thread";
+    state.activeModel = "gpt-5.5";
+    // Codex app-server 0.134.0 advertises Fast as id "priority" and reports that id as the effective service tier.
+    state.availableModels = [modelFixture("gpt-5.5", "priority")];
+    const store = createChatStateStore(state);
+    const client = clientFixture();
+    const messages: string[] = [];
+    const controller = runtimeControllerFixture(store, client, messages);
+
+    await controller.toggleFastMode();
+
+    expect(client.updateThreadSettings).toHaveBeenLastCalledWith("thread", { serviceTier: "priority" });
+    expect(store.getState().activeServiceTier).toBe("priority");
+
+    store.dispatch({ type: "thread/settings-applied", ...threadSettings("priority") });
+    await controller.toggleFastMode();
+
+    expect(client.updateThreadSettings).toHaveBeenLastCalledWith("thread", { serviceTier: null });
+    expect(store.getState().activeServiceTier).toBeNull();
+    expect(messages).toEqual(["Fast mode on for subsequent turns.", "Fast mode off for subsequent turns."]);
+  });
+
   it("leaves pending override in place when the app-server update fails", async () => {
     const state = createChatState();
     state.activeThreadId = "thread";
@@ -122,3 +147,37 @@ function clientFixture(
     ...overrides,
   };
 }
+
+function modelFixture(model: string, fastTierId: string): Model {
+  return {
+    id: model,
+    model,
+    upgrade: null,
+    upgradeInfo: null,
+    availabilityNux: null,
+    displayName: model,
+    description: "",
+    hidden: false,
+    supportedReasoningEfforts: [],
+    defaultReasoningEffort: "medium",
+    inputModalities: [],
+    supportsPersonality: true,
+    additionalSpeedTiers: ["fast"],
+    serviceTiers: [{ id: fastTierId, name: "Fast", description: "" }],
+    defaultServiceTier: null,
+    isDefault: true,
+  };
+}
+
+function threadSettings(serviceTier: string | null): Omit<Extract<ChatAction, { type: "thread/settings-applied" }>, "type"> {
+  return {
+    cwd: "/vault",
+    model: "gpt-5.5",
+    reasoningEffort: "high",
+    collaborationMode: "default",
+    serviceTier,
+    approvalPolicy: "on-request",
+    approvalsReviewer: "user",
+    activePermissionProfile: null,
+  };
+}
diff --git a/tests/features/chat/view-model.test.ts b/tests/features/chat/view-model.test.ts
index 8c1eafd8..6b2d6bad 100644
--- a/tests/features/chat/view-model.test.ts
+++ b/tests/features/chat/view-model.test.ts
@@ -51,6 +51,30 @@ describe("chat view model", () => {
     expect(model.modelChoices).toHaveLength(1);
   });
 
+  it("marks fast active for the catalog Fast service tier id", () => {
+    const state = createChatState();
+    state.activeModel = "gpt-5.5";
+    state.activeServiceTier = "priority";
+    state.effectiveConfig = effectiveConfigFixture({ model: "gpt-5.5", model_reasoning_effort: "high" });
+    state.availableModels = [modelFixture("gpt-5.5", "priority")];
+
+    const model = toolbarViewModel({
+      state,
+      snapshot: runtimeSnapshotForChatState({ state }),
+      connected: true,
+      turnBusy: false,
+      vaultPath: "/vault",
+      configuredCommand: "codex",
+      archiveConfirmThreadId: null,
+      archiveExportEnabled: true,
+      modelChoices: [],
+      effortChoices: [],
+      renameState: () => null,
+    });
+
+    expect(model.fastActive).toBe(true);
+  });
+
   it("builds slash-command status lines from chat state", () => {
     const state = createChatState();
     state.activeThreadId = "thread-1";
@@ -147,7 +171,7 @@ function threadFixture(id: string, name: string | null): Thread {
   };
 }
 
-function modelFixture(model: string): Model {
+function modelFixture(model: string, fastTierId?: string): Model {
   return {
     id: model,
     model,
@@ -161,8 +185,8 @@ function modelFixture(model: string): Model {
     defaultReasoningEffort: "high",
     inputModalities: [],
     supportsPersonality: false,
-    additionalSpeedTiers: [],
-    serviceTiers: [],
+    additionalSpeedTiers: fastTierId ? ["fast"] : [],
+    serviceTiers: fastTierId ? [{ id: fastTierId, name: "Fast", description: "" }] : [],
     defaultServiceTier: null,
     isDefault: true,
   };
diff --git a/tests/runtime/runtime-settings.test.ts b/tests/runtime/runtime-settings.test.ts
index 173d923e..e5a3be02 100644
--- a/tests/runtime/runtime-settings.test.ts
+++ b/tests/runtime/runtime-settings.test.ts
@@ -19,6 +19,8 @@ import {
   currentModel,
   currentReasoningEffort,
   currentServiceTier,
+  fastModeActive,
+  fastServiceTierRequestValue,
   fastModeLabel,
   requestedOrConfiguredServiceTier,
   requestedTurnRuntimeSettings,
@@ -241,6 +243,40 @@ describe("runtime settings", () => {
     expect(fastModeLabel(snapshot)).toBe("off");
   });
 
+  it("treats the catalog Fast service tier id as fast mode while preserving the id", () => {
+    const model = modelFixture("gpt-5.5");
+    // This mirrors Codex app-server 0.134.0 model/list: Fast is named "Fast" but its id is "priority".
+    model.serviceTiers = [{ id: "priority", name: "Fast", description: "1.5x speed, increased usage" }];
+    const snapshot = runtimeSnapshot({
+      activeModel: "gpt-5.5",
+      activeServiceTier: "priority",
+      effectiveConfig: effectiveConfigFixture({ model: "gpt-5.5" }),
+      availableModels: [model],
+    });
+
+    expect(currentServiceTier(snapshot)).toBe("priority");
+    expect(serviceTierLabel(snapshot)).toBe("priority");
+    expect(fastModeActive(snapshot)).toBe(true);
+    expect(fastModeLabel(snapshot)).toBe("on");
+    expect(fastServiceTierRequestValue(snapshot)).toBe("priority");
+  });
+
+  it("treats the Codex 0.134.0 reported default tier after clearing Fast as fast mode off", () => {
+    const model = modelFixture("gpt-5.5");
+    model.serviceTiers = [{ id: "priority", name: "Fast", description: "1.5x speed, increased usage" }];
+    const snapshot = runtimeSnapshot({
+      activeModel: "gpt-5.5",
+      activeServiceTier: "default",
+      effectiveConfig: effectiveConfigFixture({ model: "gpt-5.5" }),
+      availableModels: [model],
+    });
+
+    expect(currentServiceTier(snapshot)).toBe("default");
+    expect(serviceTierLabel(snapshot)).toBe("default");
+    expect(fastModeActive(snapshot)).toBe(false);
+    expect(fastModeLabel(snapshot)).toBe("off");
+  });
+
   it("uses requested service tier above active and configured service tiers", () => {
     const snapshot = runtimeSnapshot({
       requestedServiceTier: "off",
@@ -494,6 +530,19 @@ describe("runtime settings", () => {
     expect(requestedOrConfiguredServiceTier(snapshot)).toBeNull();
   });
 
+  it("serializes requested fast mode using the catalog Fast service tier id", () => {
+    const model = modelFixture("gpt-5.5");
+    model.serviceTiers = [{ id: "priority", name: "Fast", description: "1.5x speed, increased usage" }];
+    const snapshot = runtimeSnapshot({
+      requestedServiceTier: "fast",
+      effectiveConfig: effectiveConfigFixture({ model: "gpt-5.5" }),
+      availableModels: [model],
+    });
+
+    expect(fastModeLabel(snapshot)).toBe("on");
+    expect(requestedOrConfiguredServiceTier(snapshot)).toBe("priority");
+  });
+
   it("omits service tier when neither config nor override selects one", () => {
     expect(requestedOrConfiguredServiceTier(runtimeSnapshot({ effectiveConfig: effectiveConfigFixture({}) }))).toBeUndefined();
   });