diff --git a/fusion_accounting_assets/tests/__init__.py b/fusion_accounting_assets/tests/__init__.py
index f21110c5..e45db5b4 100644
--- a/fusion_accounting_assets/tests/__init__.py
+++ b/fusion_accounting_assets/tests/__init__.py
@@ -18,3 +18,4 @@ from . import test_assets_cron
 from . import test_engine_property
 from . import test_method_integration
 from . import test_asset_book_values_mv
+from . import test_performance_benchmarks
diff --git a/fusion_accounting_assets/tests/test_performance_benchmarks.py b/fusion_accounting_assets/tests/test_performance_benchmarks.py
new file mode 100644
index 00000000..5eb267e0
--- /dev/null
+++ b/fusion_accounting_assets/tests/test_performance_benchmarks.py
@@ -0,0 +1,117 @@
+"""Performance benchmarks tagged 'benchmark'."""
+
+import json
+import statistics
+import time
+from datetime import date
+
+from odoo.tests.common import HttpCase, TransactionCase, new_test_user
+from odoo.tests import tagged
+
+
+@tagged('post_install', '-at_install', 'benchmark')
+class TestEngineBenchmarks(TransactionCase):
+
+    def setUp(self):
+        super().setUp()
+        self.engine = self.env['fusion.asset.engine']
+
+    def _percentile(self, samples, p):
+        if len(samples) <= 1:
+            return samples[0] if samples else 0
+        sorted_s = sorted(samples)
+        idx = int(len(sorted_s) * p / 100)
+        return sorted_s[min(idx, len(sorted_s) - 1)]
+
+    def test_compute_schedule_p95(self):
+        timings = []
+        for i in range(10):
+            asset = self.env['fusion.asset'].create({
+                'name': f'PerfAsset{i}', 'cost': 100000, 'salvage_value': 5000,
+                'acquisition_date': date(2026, 1, 1),
+                'method': 'straight_line', 'useful_life_years': 10,
+            })
+            start = time.perf_counter()
+            self.engine.compute_depreciation_schedule(asset)
+            timings.append((time.perf_counter() - start) * 1000)
+        p95 = self._percentile(timings, 95)
+        median = statistics.median(timings)
+        msg = f"compute_schedule(10yr): median={median:.0f}ms p95={p95:.0f}ms"
+        print(f"\n  PERF: {msg} (target <500ms)")
+        self.assertLess(p95, 5000, f"way over budget: {msg}")
+
+    def test_post_depreciation_p95(self):
+        asset = self.env['fusion.asset'].create({
+            'name': 'PostPerf', 'cost': 50000,
+            'acquisition_date': date(2026, 1, 1),
+            'in_service_date': date(2026, 1, 1),
+            'method': 'straight_line', 'useful_life_years': 10,
+        })
+        self.engine.compute_depreciation_schedule(asset)
+        asset.action_set_running()
+        timings = []
+        for _ in range(5):
+            start = time.perf_counter()
+            self.engine.post_depreciation_entry(asset)
+            timings.append((time.perf_counter() - start) * 1000)
+        p95 = self._percentile(timings, 95)
+        median = statistics.median(timings)
+        msg = f"post_depreciation: median={median:.0f}ms p95={p95:.0f}ms"
+        print(f"\n  PERF: {msg} (target <300ms)")
+        self.assertLess(p95, 3000)
+
+    def test_dispose_asset_p95(self):
+        timings = []
+        for i in range(5):
+            asset = self.env['fusion.asset'].create({
+                'name': f'DispPerf{i}', 'cost': 10000,
+                'acquisition_date': date(2026, 1, 1),
+                'in_service_date': date(2026, 1, 1),
+                'method': 'straight_line', 'useful_life_years': 5,
+            })
+            self.engine.compute_depreciation_schedule(asset)
+            asset.action_set_running()
+            start = time.perf_counter()
+            self.engine.dispose_asset(asset, sale_amount=5000)
+            timings.append((time.perf_counter() - start) * 1000)
+        p95 = self._percentile(timings, 95)
+        median = statistics.median(timings)
+        msg = f"dispose_asset: median={median:.0f}ms p95={p95:.0f}ms"
+        print(f"\n  PERF: {msg} (target <300ms)")
+        self.assertLess(p95, 3000)
+
+
+@tagged('post_install', '-at_install', 'benchmark')
+class TestControllerBenchmarks(HttpCase):
+
+    def test_list_endpoint_p95(self):
+        new_test_user(
+            self.env, login='asset_perf',
+            groups='base.group_user,account.group_account_invoice',
+        )
+        for i in range(20):
+            self.env['fusion.asset'].create({
+                'name': f'ListPerf{i}', 'cost': 1000,
+                'acquisition_date': date(2026, 1, 1),
+                'method': 'straight_line', 'useful_life_years': 4,
+            })
+        self.authenticate('asset_perf', 'asset_perf')
+        timings = []
+        for _ in range(5):
+            start = time.perf_counter()
+            response = self.url_open(
+                '/fusion/assets/list',
+                data=json.dumps({
+                    'jsonrpc': '2.0', 'method': 'call', 'id': 1,
+                    'params': {'company_id': self.env.company.id},
+                }),
+                headers={'Content-Type': 'application/json'},
+            )
+            timings.append((time.perf_counter() - start) * 1000)
+            self.assertEqual(response.status_code, 200)
+        sorted_t = sorted(timings)
+        p95 = sorted_t[min(int(len(sorted_t) * 0.95), len(sorted_t) - 1)]
+        median = statistics.median(timings)
+        msg = f"controller.list: median={median:.0f}ms p95={p95:.0f}ms"
+        print(f"\n  PERF: {msg} (target <300ms)")
+        self.assertLess(p95, 3000)