diff --git a/autogpt_platform/backend/backend/data/platform_cost.py b/autogpt_platform/backend/backend/data/platform_cost.py
index 443790ac5d..11ed885b4a 100644
--- a/autogpt_platform/backend/backend/data/platform_cost.py
+++ b/autogpt_platform/backend/backend/data/platform_cost.py
@@ -3,12 +3,11 @@ import logging
 from datetime import datetime, timedelta, timezone
 from typing import Any
 
-from typing_extensions import TypedDict
-
 from prisma.models import PlatformCostLog as PrismaLog
 from prisma.models import User as PrismaUser
 from prisma.types import PlatformCostLogCreateInput, PlatformCostLogWhereInput
 from pydantic import BaseModel
+from typing_extensions import TypedDict
 
 from backend.data.db import query_raw_with_schema
 from backend.util.cache import cached
@@ -143,6 +142,7 @@ class UserCostSummary(BaseModel):
     total_input_tokens: int
     total_output_tokens: int
     request_count: int
+    cost_bearing_request_count: int = 0
 
 
 class CostLogRow(BaseModel):
@@ -286,7 +286,7 @@ async def get_platform_cost_dashboard(
     # queries so they honour all active dashboard filters, not just start date.
     raw_params: list = [start]
     raw_where_clauses = [
-        '"trackingType" = \'cost_usd\'',
+        "\"trackingType\" = 'cost_usd'",
         '"createdAt" >= $1',
     ]
     param_idx = 2  # $1 is already start
@@ -316,12 +316,20 @@ async def get_platform_cost_dashboard(
         raw_params.append(block_name)
         param_idx += 1
 
+    # If the caller supplied a specific tracking_type filter, replace the
+    # hardcoded cost_usd clause so the percentile/bucket queries respect it.
+    if tracking_type is not None:
+        raw_where_clauses[0] = f'"trackingType" = ${param_idx}'
+        raw_params.append(tracking_type)
+        param_idx += 1
+
     raw_where = " AND ".join(raw_where_clauses)
 
-    # Run all six aggregation queries in parallel.
+    # Run all seven aggregation queries in parallel.
     (
         by_provider_groups,
         by_user_groups,
+        by_user_tracking_groups,
         total_user_groups,
         total_agg_groups,
         percentile_rows,
@@ -342,6 +350,13 @@ async def get_platform_cost_dashboard(
             sum=sum_fields,
             count=True,
         ),
+        # Per-user cost-bearing request count: group by (userId, trackingType)
+        # so we can compute the correct denominator for per-user avg cost.
+        PrismaLog.prisma().group_by(
+            by=["userId", "trackingType"],
+            where=where,
+            count=True,
+        ),
         # Distinct user count: group by userId, count groups.
         PrismaLog.prisma().group_by(
             by=["userId"],
@@ -376,6 +391,8 @@ async def get_platform_cost_dashboard(
             *raw_params,
         ),
         # Histogram buckets for cost distribution (respects all filters).
+        # NULL costMicrodollars is excluded explicitly to prevent such rows
+        # from falling through all WHEN clauses into the ELSE '$10+' bucket.
         query_raw_with_schema(
             "SELECT"
             "  CASE"
@@ -393,7 +410,7 @@ async def get_platform_cost_dashboard(
             "  END as bucket,"
             "  COUNT(*) as count"
             ' FROM {schema_prefix}"PlatformCostLog"'
-            f" WHERE {raw_where}"
+            f' WHERE {raw_where} AND "costMicrodollars" IS NOT NULL'
             " GROUP BY bucket"
             ' ORDER BY MIN("costMicrodollars")',
             *raw_params,
@@ -448,6 +465,16 @@ async def get_platform_cost_dashboard(
         _ca(r) for r in total_agg_groups if r.get("trackingType") == "tokens"
     )
 
+    # Per-user cost-bearing request count: used for per-user avg cost so the
+    # denominator matches the numerator (cost_usd rows only, per user).
+    user_cost_bearing_counts: dict[str, int] = {}
+    for r in by_user_tracking_groups:
+        if r.get("trackingType") == "cost_usd" and r.get("userId"):
+            uid = r["userId"]
+            user_cost_bearing_counts[uid] = user_cost_bearing_counts.get(uid, 0) + _ca(
+                r
+            )
+
     return PlatformCostDashboard(
         by_provider=[
             ProviderCostSummary(
@@ -473,6 +500,9 @@ async def get_platform_cost_dashboard(
                 total_input_tokens=_si(r, "inputTokens"),
                 total_output_tokens=_si(r, "outputTokens"),
                 request_count=_ca(r),
+                cost_bearing_request_count=user_cost_bearing_counts.get(
+                    r.get("userId") or "", 0
+                ),
             )
             for r in by_user_groups
         ],
diff --git a/autogpt_platform/backend/backend/data/platform_cost_test.py b/autogpt_platform/backend/backend/data/platform_cost_test.py
index 92470d0a64..5144d29fbb 100644
--- a/autogpt_platform/backend/backend/data/platform_cost_test.py
+++ b/autogpt_platform/backend/backend/data/platform_cost_test.py
@@ -286,6 +286,7 @@ class TestGetPlatformCostDashboard:
             side_effect=[
                 [provider_row],  # by_provider
                 [user_row],  # by_user
+                [],  # by_user_tracking_groups (no cost_usd rows for this user)
                 [{"userId": "u1"}],  # distinct users
                 [provider_row],  # total agg
             ]
@@ -322,7 +323,9 @@ class TestGetPlatformCostDashboard:
         assert len(dashboard.by_user) == 1
         assert dashboard.by_user[0].email == "a***@b.com"
         assert dashboard.cost_p50_microdollars == 1000
+        assert dashboard.cost_p75_microdollars == 2000
         assert dashboard.cost_p95_microdollars == 4000
+        assert dashboard.cost_p99_microdollars == 5000
         assert len(dashboard.cost_buckets) == 1
         # Token averages must use token_bearing_requests (3) not cost_bearing (0)
         assert dashboard.avg_input_tokens_per_request == pytest.approx(1000 / 3)
@@ -351,6 +354,7 @@ class TestGetPlatformCostDashboard:
             side_effect=[
                 [provider_row],  # by_provider
                 [user_row],  # by_user
+                [],  # by_user_tracking_groups
                 [{"userId": "u2"}],  # distinct users
                 [provider_row],  # total agg
             ]
@@ -385,7 +389,7 @@ class TestGetPlatformCostDashboard:
     @pytest.mark.asyncio
     async def test_returns_empty_dashboard(self):
         mock_actions = MagicMock()
-        mock_actions.group_by = AsyncMock(side_effect=[[], [], [], []])
+        mock_actions.group_by = AsyncMock(side_effect=[[], [], [], [], []])
         mock_actions.find_many = AsyncMock(return_value=[])
 
         with (
@@ -418,7 +422,7 @@ class TestGetPlatformCostDashboard:
         start = datetime(2026, 1, 1, tzinfo=timezone.utc)
 
         mock_actions = MagicMock()
-        mock_actions.group_by = AsyncMock(side_effect=[[], [], [], []])
+        mock_actions.group_by = AsyncMock(side_effect=[[], [], [], [], []])
         mock_actions.find_many = AsyncMock(return_value=[])
 
         raw_mock = AsyncMock(side_effect=[[], []])
@@ -440,8 +444,8 @@ class TestGetPlatformCostDashboard:
                 start=start, provider="openai", user_id="u1"
             )
 
-        # group_by called 4 times (by_provider, by_user, distinct users, totals)
-        assert mock_actions.group_by.await_count == 4
+        # group_by called 5 times (by_provider, by_user, by_user_tracking, distinct users, totals)
+        assert mock_actions.group_by.await_count == 5
         # The where dict passed to the first call should include createdAt
         first_call_kwargs = mock_actions.group_by.call_args_list[0][1]
         assert "createdAt" in first_call_kwargs.get("where", {})
diff --git a/autogpt_platform/frontend/src/app/(platform)/admin/platform-costs/components/UserTable.tsx b/autogpt_platform/frontend/src/app/(platform)/admin/platform-costs/components/UserTable.tsx
index ed3116381a..a6d54dbdea 100644
--- a/autogpt_platform/frontend/src/app/(platform)/admin/platform-costs/components/UserTable.tsx
+++ b/autogpt_platform/frontend/src/app/(platform)/admin/platform-costs/components/UserTable.tsx
@@ -58,9 +58,11 @@ function UserTable({ data }: Props) {
                 {formatTokens(row.total_output_tokens)}
               </td>
               <td className="px-4 py-3 text-right">
-                {row.request_count > 0 && row.total_cost_microdollars > 0
+                {(row.cost_bearing_request_count ?? 0) > 0 &&
+                row.total_cost_microdollars > 0
                   ? formatMicrodollars(
-                      row.total_cost_microdollars / row.request_count,
+                      row.total_cost_microdollars /
+                        row.cost_bearing_request_count!,
                     )
                   : "-"}
               </td>
diff --git a/autogpt_platform/frontend/src/app/api/__generated__/models/userCostSummary.ts b/autogpt_platform/frontend/src/app/api/__generated__/models/userCostSummary.ts
new file mode 100644
index 0000000000..ded1a5fe69
--- /dev/null
+++ b/autogpt_platform/frontend/src/app/api/__generated__/models/userCostSummary.ts
@@ -0,0 +1,19 @@
+/**
+ * Generated by orval v7.13.0 🍺
+ * Do not edit manually.
+ * AutoGPT Agent Server
+ * This server is used to execute agents that are created by the AutoGPT system.
+ * OpenAPI spec version: 0.1
+ */
+import type { UserCostSummaryUserId } from "./userCostSummaryUserId";
+import type { UserCostSummaryEmail } from "./userCostSummaryEmail";
+
+export interface UserCostSummary {
+  user_id?: UserCostSummaryUserId;
+  email?: UserCostSummaryEmail;
+  total_cost_microdollars: number;
+  total_input_tokens: number;
+  total_output_tokens: number;
+  request_count: number;
+  cost_bearing_request_count?: number;
+}
diff --git a/autogpt_platform/frontend/src/app/api/openapi.json b/autogpt_platform/frontend/src/app/api/openapi.json
index 19e2c86567..ca9eee94d6 100644
--- a/autogpt_platform/frontend/src/app/api/openapi.json
+++ b/autogpt_platform/frontend/src/app/api/openapi.json
@@ -15585,7 +15585,8 @@
             "type": "integer",
             "title": "Total Output Tokens"
           },
-          "request_count": { "type": "integer", "title": "Request Count" }
+          "request_count": { "type": "integer", "title": "Request Count" },
+          "cost_bearing_request_count": { "type": "integer", "title": "Cost Bearing Request Count", "default": 0 }
         },
         "type": "object",
         "required": [