fix: force cost_usd for percentile/histogram queries, fix test + prettier

- Backend: always pass tracking_type=None to _build_raw_where for
  percentile and histogram queries so they compute stats on cost_usd
  rows regardless of the caller's tracking_type filter.
- Frontend test: use getAllByText for "5" which appears in both the
  Active Users card and the $1-2 bucket count.
- Frontend: fix prettier formatting in PlatformCostContent.tsx.

autogpt_platform/backend/backend/data/platform_cost.py

@@ -352,8 +352,11 @@ async def get_platform_cost_dashboard(
     # Build parameterised WHERE clause for the raw SQL percentile/bucket
     # queries.  Uses _build_raw_where so filter logic is shared with
     # _build_prisma_where and only maintained in one place.
+    # Always force tracking_type=None here so _build_raw_where defaults to
+    # "cost_usd" — percentile and histogram queries only make sense on
+    # cost-denominated rows, regardless of what the caller is filtering.
     raw_where, raw_params = _build_raw_where(
-        start, end, provider, user_id, model, block_name, tracking_type
+        start, end, provider, user_id, model, block_name, tracking_type=None
     )
 
     # Queries that always run regardless of tracking_type filter.

autogpt_platform/frontend/src/app/(platform)/admin/platform-costs/__tests__/PlatformCostContent.test.tsx

@@ -185,7 +185,9 @@ describe("PlatformCostContent", () => {
     );
     expect(screen.getByText("$5.0000")).toBeDefined();
     expect(screen.getByText("100")).toBeDefined();
-    expect(screen.getByText("5")).toBeDefined();
+    // "5" appears in multiple places (Active Users card + bucket count),
+    // so verify at least one element renders it.
+    expect(screen.getAllByText("5").length).toBeGreaterThanOrEqual(1);
     expect(screen.getByText("openai")).toBeDefined();
     expect(screen.getByText("google_maps")).toBeDefined();
   });

autogpt_platform/frontend/src/app/(platform)/admin/platform-costs/components/PlatformCostContent.tsx

@@ -264,8 +264,7 @@ export function PlatformCostContent({ searchParams }: Props) {
                     value: Math.round(
                       dashboard.avg_output_tokens_per_request ?? 0,
                     ).toLocaleString(),
-                    subtitle:
-                      "Completion tokens per request (response length)",
+                    subtitle: "Completion tokens per request (response length)",
                   },
                   {
                     label: "Total Tokens",