diff --git a/changelog.d/maintenance/release-v3850-basereds-error-helper-20260819.md b/changelog.d/maintenance/release-v3850-basereds-error-helper-20260819.md
new file mode 100644
index 0000000000..e89d53ab49
--- /dev/null
+++ b/changelog.d/maintenance/release-v3850-basereds-error-helper-20260819.md
@@ -0,0 +1,3 @@
+- **fix(ci):** route `open-sse/handlers/imageGeneration/providers/geminiWeb.ts`'s b64_json
+  download-failure message through `sanitizeErrorMessage()` instead of embedding a raw
+  `err.message`, clearing the `check:error-helper` base-red on `release/v3.8.50` (#9985).
diff --git a/changelog.d/maintenance/release-v3850-basereds-testdrift-20260819.md b/changelog.d/maintenance/release-v3850-basereds-testdrift-20260819.md
new file mode 100644
index 0000000000..f4deae6aae
--- /dev/null
+++ b/changelog.d/maintenance/release-v3850-basereds-testdrift-20260819.md
@@ -0,0 +1,12 @@
+- **fix(tests):** drain several base-reds on `release/v3.8.50` (#9985) that were all instances
+  of the same pattern — a legitimate product change landed without updating the test that
+  asserted the old behavior: `tests/unit/glm-provider-model-import-route.test.ts` (12 tests)
+  and `tests/unit/model-sync-route.test.ts` (2 tests) predate #10603's "upstream model sync is
+  opt-in and manual overrides are preserved" change; `tests/unit/antigravity-model-aliases.test.ts`
+  predated #10537 retiring the collapsed `gemini-3.7-flash` alias in favor of its three tiered
+  ids. Also fixes a real data drift in `open-sse/config/freeModelCatalog.data.ts` (the `qwen-web`
+  free-catalog entry still pointed at the retired `qwen3.8-max-preview` id instead of the
+  current `qwen3.8-max`), corrects the zh-TW `providers.autoFetchModelsTooltip` string to the
+  glossary-canonical 快取 instead of 緩存, and removes an unused default export from
+  `src/lib/oauth/providers/zed-hosted.ts` (the named export already covers every consumer) to
+  shave one symbol off the `check:dead-code` ratchet regression.
diff --git a/open-sse/config/freeModelCatalog.data.ts b/open-sse/config/freeModelCatalog.data.ts
index decef67c50..3d1777ed65 100644
--- a/open-sse/config/freeModelCatalog.data.ts
+++ b/open-sse/config/freeModelCatalog.data.ts
@@ -361,7 +361,7 @@ export const FREE_MODEL_BUDGETS: FreeModelBudget[] = [
   { provider: "qoder", modelId: "deepseek-v4-pro", displayName: "DeepSeek-V4-Pro", monthlyTokens: 0, creditTokens: 1000000, freeType: "one-time-initial", poolKey: "qoder", tos: "caution" },
   { provider: "qoder", modelId: "deepseek-v4-flash", displayName: "DeepSeek-V4-Flash", monthlyTokens: 0, creditTokens: 1000000, freeType: "one-time-initial", poolKey: "qoder", tos: "caution" },
   { provider: "qoder", modelId: "minimax-m3", displayName: "MiniMax-M3", monthlyTokens: 0, creditTokens: 1000000, freeType: "one-time-initial", poolKey: "qoder", tos: "caution" },
-  { provider: "qwen-web", modelId: "qwen3.8-max-preview", displayName: "Qwen3.8 Max Preview", monthlyTokens: 0, creditTokens: 0, freeType: "keyless", poolKey: "qwen-web", tos: "avoid" },
+  { provider: "qwen-web", modelId: "qwen3.8-max", displayName: "Qwen3.8 Max", monthlyTokens: 0, creditTokens: 0, freeType: "keyless", poolKey: "qwen-web", tos: "avoid" },
   { provider: "qwen-web", modelId: "qwen3.7-max", displayName: "Qwen3.7 Max", monthlyTokens: 0, creditTokens: 0, freeType: "keyless", poolKey: "qwen-web", tos: "avoid" },
   { provider: "qwen-web", modelId: "qwen3.7-plus", displayName: "Qwen3.7 Plus", monthlyTokens: 0, creditTokens: 0, freeType: "keyless", poolKey: "qwen-web", tos: "avoid" },
   { provider: "qwen-web", modelId: "qwen3.6-plus", displayName: "Qwen3.6 Plus", monthlyTokens: 0, creditTokens: 0, freeType: "keyless", poolKey: "qwen-web", tos: "avoid" },
diff --git a/open-sse/handlers/imageGeneration/providers/geminiWeb.ts b/open-sse/handlers/imageGeneration/providers/geminiWeb.ts
index 43c91de8fc..8f131fe83d 100644
--- a/open-sse/handlers/imageGeneration/providers/geminiWeb.ts
+++ b/open-sse/handlers/imageGeneration/providers/geminiWeb.ts
@@ -20,6 +20,7 @@
 import { GeminiWebExecutor } from "../../../executors/gemini-web.ts";
 import { fetchRemoteImage } from "@/shared/network/remoteImageFetch";
 import { saveImageErrorResult, saveImageSuccessResult } from "../../imageGeneration.ts";
+import { sanitizeErrorMessage } from "../../../utils/error.ts";
 
 /** Each image is one gemini.google.com turn (~30-60s). Cap like chatgpt-web. */
 const GEMINI_WEB_IMAGE_N_MAX = 4;
@@ -210,7 +211,7 @@ export async function handleGeminiWebImageGeneration({
           model,
           status: 502,
           startTime,
-          error: `Gemini Web generated an image but OmniRoute could not download it for b64_json conversion: ${err instanceof Error ? err.message : String(err)}`,
+          error: `Gemini Web generated an image but OmniRoute could not download it for b64_json conversion: ${sanitizeErrorMessage(err instanceof Error ? err.message : String(err))}`,
           requestBody,
         });
       }
diff --git a/src/i18n/messages/zh-TW.json b/src/i18n/messages/zh-TW.json
index 627b7b4383..a7ac344ae8 100644
--- a/src/i18n/messages/zh-TW.json
+++ b/src/i18n/messages/zh-TW.json
@@ -6350,7 +6350,7 @@
     "autoFetchModels": "自動獲取上游模型",
     "autoFetchModelsDisabled": "上游模型自動獲取已禁用",
     "autoFetchModelsEnabled": "上游模型自動獲取已啟用",
-    "autoFetchModelsTooltip": "在需要時獲取並緩存上游模型",
+    "autoFetchModelsTooltip": "在需要時獲取並快取上游模型",
     "autoFetchModelsToggleFailed": "無法切換上游模型自動獲取",
     "autoFetchModelsPartialFailure": "某些連接已更新，但上游模型自動獲取並未在所有地方更改",
     "overridesUpstreamModel": "覆蓋上游",
diff --git a/src/lib/oauth/providers/zed-hosted.ts b/src/lib/oauth/providers/zed-hosted.ts
index 46a4c1b320..b37c157163 100644
--- a/src/lib/oauth/providers/zed-hosted.ts
+++ b/src/lib/oauth/providers/zed-hosted.ts
@@ -126,5 +126,3 @@ export const zedHosted = {
     },
   }),
 };
-
-export default zedHosted;
diff --git a/tests/unit/antigravity-model-aliases.test.ts b/tests/unit/antigravity-model-aliases.test.ts
index c821766058..542573b3a4 100644
--- a/tests/unit/antigravity-model-aliases.test.ts
+++ b/tests/unit/antigravity-model-aliases.test.ts
@@ -17,8 +17,11 @@ function getPublicModel(id: string) {
   return ANTIGRAVITY_PUBLIC_MODELS.find((model) => model.id === id) as any;
 }
 
+// #10537 retired the single-alias `gemini-3.7-flash` (which mapped to the upstream
+// `gemini-3.7-flash-tiered`) in favor of three directly-callable tiered public models —
+// the suffixed ids now work upstream without the collapsing alias. Keep this list in sync
+// with ANTIGRAVITY_PUBLIC_MODELS/ANTIGRAVITY_MODEL_ALIASES instead of the retired bare id.
 const EXPECTED_FLASH_TIERS = [
-  ["gemini-3.7-flash", "Gemini 3.7 Flash"],
   ["gemini-3.7-flash-high", "Gemini 3.7 Flash (High)"],
   ["gemini-3.7-flash-medium", "Gemini 3.7 Flash (Medium)"],
   ["gemini-3.7-flash-low", "Gemini 3.7 Flash (Low)"],
diff --git a/tests/unit/glm-provider-model-import-route.test.ts b/tests/unit/glm-provider-model-import-route.test.ts
index dd174c93c6..b86c115e70 100644
--- a/tests/unit/glm-provider-model-import-route.test.ts
+++ b/tests/unit/glm-provider-model-import-route.test.ts
@@ -4,6 +4,12 @@ import fs from "node:fs";
 import os from "node:os";
 import path from "node:path";
 
+// #10603 made upstream model sync opt-in (isAutoFetchModelsEnabled() now requires
+// providerSpecificData.autoFetchModels === true) so remote discovery doesn't overwrite
+// manual catalog overrides by default. Every connection below sets it explicitly so the
+// mocked `fetch` in each test actually gets called — without it, the route short-circuits
+// to the local/cached catalog before ever reaching the network call these tests assert on.
+
 const TEST_DATA_DIR = fs.mkdtempSync(path.join(os.tmpdir(), "omniroute-glm-models-"));
 process.env.DATA_DIR = TEST_DATA_DIR;
 
@@ -29,7 +35,7 @@ test("GLM import uses international coding endpoint when apiRegion is internatio
     authType: "apikey",
     name: "glm-intl",
     apiKey: "glm-key",
-    providerSpecificData: { apiRegion: "international" },
+    providerSpecificData: { apiRegion: "international", autoFetchModels: true },
   });
 
   const originalFetch = globalThis.fetch;
@@ -88,7 +94,7 @@ test("GLM import normalizes custom coding models URLs without duplicating endpoi
         authType: "apikey",
         name: `glm-custom-${index}`,
         apiKey: testCase.apiKey,
-        providerSpecificData: { baseUrl: testCase.baseUrl },
+        providerSpecificData: { baseUrl: testCase.baseUrl, autoFetchModels: true },
       })
     );
   }
@@ -129,7 +135,7 @@ test("GLM import falls back to Anthropic model discovery when coding discovery f
     authType: "apikey",
     name: "glm-discovery-fallback",
     apiKey: "glm-key",
-    providerSpecificData: { apiRegion: "international" },
+    providerSpecificData: { apiRegion: "international", autoFetchModels: true },
   });
 
   const originalFetch = globalThis.fetch;
@@ -178,7 +184,7 @@ test("GLM import preserves auth failures instead of falling back across transpor
     authType: "apikey",
     name: "glm-auth-fail",
     apiKey: "bad-key",
-    providerSpecificData: { apiRegion: "international" },
+    providerSpecificData: { apiRegion: "international", autoFetchModels: true },
   });
 
   const originalFetch = globalThis.fetch;
@@ -207,7 +213,7 @@ test("GLMT import shares the GLM coding models endpoint and surfaces provider me
     authType: "apikey",
     name: "glmt-intl",
     apiKey: "glmt-key",
-    providerSpecificData: { apiRegion: "international" },
+    providerSpecificData: { apiRegion: "international", autoFetchModels: true },
   });
 
   const originalFetch = globalThis.fetch;
@@ -241,7 +247,7 @@ test("GLM import uses China coding endpoint when apiRegion is china", async () =
     authType: "apikey",
     name: "glm-cn",
     apiKey: "glm-cn-key",
-    providerSpecificData: { apiRegion: "china" },
+    providerSpecificData: { apiRegion: "china", autoFetchModels: true },
   });
 
   const originalFetch = globalThis.fetch;
@@ -272,7 +278,7 @@ test("GLM China provider import uses the specialized GLM discovery path", async
     authType: "apikey",
     name: "glm-cn-provider",
     apiKey: "glm-cn-key",
-    providerSpecificData: {},
+    providerSpecificData: { autoFetchModels: true },
   });
 
   const originalFetch = globalThis.fetch;
@@ -304,7 +310,7 @@ test("GLM import defaults to international endpoint when apiRegion is missing",
     authType: "apikey",
     name: "glm-default",
     apiKey: "glm-key",
-    providerSpecificData: {},
+    providerSpecificData: { autoFetchModels: true },
   });
 
   const originalFetch = globalThis.fetch;
@@ -331,7 +337,7 @@ test("GLM import defaults to international endpoint when apiRegion is invalid",
     authType: "apikey",
     name: "glm-bogus",
     apiKey: "glm-key",
-    providerSpecificData: { apiRegion: "bogus" },
+    providerSpecificData: { apiRegion: "bogus", autoFetchModels: true },
   });
 
   const originalFetch = globalThis.fetch;
@@ -359,7 +365,7 @@ test("GLM import prefers apiKey over accessToken and sends only Authorization Be
     name: "glm-both-tokens",
     apiKey: "glm-api-key",
     accessToken: "glm-access-token",
-    providerSpecificData: { apiRegion: "international" },
+    providerSpecificData: { apiRegion: "international", autoFetchModels: true },
   });
 
   const originalFetch = globalThis.fetch;
@@ -387,7 +393,7 @@ test("GLM import falls back to accessToken when apiKey is absent", async () => {
     authType: "apikey",
     name: "glm-access-only",
     accessToken: "glm-access-token",
-    providerSpecificData: { apiRegion: "international" },
+    providerSpecificData: { apiRegion: "international", autoFetchModels: true },
   });
 
   const originalFetch = globalThis.fetch;
@@ -414,7 +420,7 @@ test("GLM import falls back to the local catalog on upstream non-OK status codes
     authType: "apikey",
     name: "glm-error",
     apiKey: "glm-key",
-    providerSpecificData: { apiRegion: "international" },
+    providerSpecificData: { apiRegion: "international", autoFetchModels: true },
   });
 
   const originalFetch = globalThis.fetch;
diff --git a/tests/unit/model-sync-route.test.ts b/tests/unit/model-sync-route.test.ts
index 0e71b96c17..ece60f730d 100644
--- a/tests/unit/model-sync-route.test.ts
+++ b/tests/unit/model-sync-route.test.ts
@@ -471,6 +471,10 @@ test("model sync route writes synced available models for non-Gemini providers t
   ]);
 });
 
+// #10603 ("make upstream model sync opt-in and preserve manual overrides") also applies to
+// import/merge mode: a manual custom-model row sharing an id with a discovered model is kept
+// as the user-owned overlay instead of being demoted/removed. See the comment above the
+// "reports synced managed models separately from preserved manual models" test.
 test("model sync route import mode merges discovered models without deleting manual models", async () => {
   await resetStorage();
 
@@ -512,10 +516,13 @@ test("model sync route import mode merges discovered models without deleting man
   assert.equal(body.updatedCount, 0);
   assert.equal(body.syncedAliases, 1);
   assert.deepEqual(body.modelChanges, { added: 1, removed: 0, updated: 0, total: 1 });
-  assert.deepEqual(body.customModelChanges, { added: 0, removed: 1, updated: 0, total: 1 });
+  assert.deepEqual(body.customModelChanges, { added: 0, removed: 0, updated: 0, total: 0 });
   assert.deepEqual(
     body.models.map((model) => ({ id: model.id, source: model.source })),
-    [{ id: "manual-only", source: "manual" }]
+    [
+      { id: "manual-only", source: "manual" },
+      { id: "router-v4", source: "manual" },
+    ]
   );
   assert.deepEqual(
     body.importedModels.map((model) => ({ id: model.id, source: model.source })),
@@ -820,6 +827,13 @@ test("model sync route forwards cookies, filters built-ins, and syncs aliases fo
   assert.equal(logs[0].account, "External Sync");
 });
 
+// #10603 ("make upstream model sync opt-in and preserve manual overrides") changed
+// importManagedModels() so a manually configured custom-model row that shares an id with
+// a synced upstream model is no longer demoted/removed — it stays as the user-owned
+// metadata overlay and is merged over the synced base at read time (see the comment in
+// src/lib/providerModels/managedModelImport.ts). Only rows already tagged as an
+// imported/auto-sync source get pruned. This test predates that change; its expectations
+// below reflect the current preserve-manual-overrides behavior.
 test("model sync route reports synced managed models separately from preserved manual models", async () => {
   await resetStorage();
 
@@ -858,10 +872,13 @@ test("model sync route reports synced managed models separately from preserved m
   assert.equal(body.availableModelsCount, 2);
   assert.equal(body.importedCount, 1);
   assert.equal(body.updatedCount, 0);
-  assert.deepEqual(body.customModelChanges, { added: 0, removed: 1, updated: 0, total: 1 });
+  assert.deepEqual(body.customModelChanges, { added: 0, removed: 0, updated: 0, total: 0 });
   assert.deepEqual(
     body.models.map((model) => ({ id: model.id, source: model.source })),
-    [{ id: "manual-only", source: "manual" }]
+    [
+      { id: "manual-only", source: "manual" },
+      { id: "router-v4", source: "manual" },
+    ]
   );
   assert.deepEqual(
     (await modelsDb.getSyncedAvailableModels("openrouter")).map((model) => ({