diff --git a/tests/tools/test_delegate_cost_footer.py b/tests/tools/test_delegate_cost_footer.py
new file mode 100644
index 0000000000000..413a6fe7c354c
--- /dev/null
+++ b/tests/tools/test_delegate_cost_footer.py
@@ -0,0 +1,118 @@
+#!/usr/bin/env python3
+"""Per-delegation cost in the serialized result entry.
+
+Each entry in the results array returned to the parent model should
+carry the child's spend (cost_usd + cost_status) so the model can see
+what each delegation cost — previously the cost was only folded into
+the parent session total and stripped from the entry.
+
+The parent rollup (session_estimated_cost_usd fold via _child_cost_usd)
+must stay unchanged, and the internal _child_cost_usd field must still
+be stripped before serialization.
+
+Inspired by: Perplexity Agent API result shape (idea-level)
+"""
+
+import json
+import threading
+import unittest
+from unittest.mock import MagicMock, patch
+
+from tools.delegate_tool import delegate_task
+
+
+def _make_mock_parent(depth=0):
+    parent = MagicMock()
+    parent.base_url = "https://openrouter.ai/api/v1"
+    parent.api_key = "test-key"
+    parent.provider = "openrouter"
+    parent.api_mode = "chat_completions"
+    parent.model = "anthropic/claude-sonnet-4"
+    parent.platform = "cli"
+    parent.providers_allowed = None
+    parent.providers_ignored = None
+    parent.providers_order = None
+    parent.provider_sort = None
+    parent._session_db = None
+    parent._delegate_depth = depth
+    parent._active_children = []
+    parent._active_children_lock = threading.Lock()
+    parent._print_fn = None
+    parent.tool_progress_callback = None
+    parent.thinking_callback = None
+    parent.session_estimated_cost_usd = 0.0
+    parent.session_cost_source = "none"
+    parent.session_cost_status = "unknown"
+    return parent
+
+
+def _make_mock_child(cost=0.1234567, cost_status="estimated"):
+    child = MagicMock()
+    child.run_conversation.return_value = {
+        "final_response": "done",
+        "completed": True,
+        "api_calls": 2,
+        "messages": [],
+    }
+    child.session_prompt_tokens = 100
+    child.session_completion_tokens = 50
+    child.session_estimated_cost_usd = cost
+    child.session_cost_status = cost_status
+    child.model = "anthropic/claude-sonnet-4"
+    child.session_id = "child-session"
+    return child
+
+
+class TestCostInResultEntry(unittest.TestCase):
+    def _run(self, child):
+        parent = _make_mock_parent()
+        with patch("run_agent.AIAgent", return_value=child):
+            result = json.loads(
+                delegate_task(goal="Test per-delegation cost", parent_agent=parent)
+            )
+        return parent, result
+
+    def test_entry_carries_cost_usd_and_status(self):
+        child = _make_mock_child(cost=0.1234567, cost_status="estimated")
+        _, result = self._run(child)
+        entry = result["results"][0]
+        self.assertIn("cost_usd", entry)
+        self.assertAlmostEqual(entry["cost_usd"], 0.123457, places=6)
+        self.assertEqual(entry["cost_status"], "estimated")
+
+    def test_reported_status_passes_through(self):
+        child = _make_mock_child(cost=0.5, cost_status="reported")
+        _, result = self._run(child)
+        entry = result["results"][0]
+        self.assertEqual(entry["cost_status"], "reported")
+
+    def test_zero_cost_child_has_zero_cost_entry(self):
+        child = _make_mock_child(cost=0.0, cost_status="unknown")
+        _, result = self._run(child)
+        entry = result["results"][0]
+        self.assertEqual(entry["cost_usd"], 0.0)
+
+    def test_internal_child_cost_field_still_stripped(self):
+        child = _make_mock_child(cost=0.25)
+        _, result = self._run(child)
+        for entry in result["results"]:
+            self.assertNotIn("_child_cost_usd", entry)
+            self.assertNotIn("_child_role", entry)
+
+    def test_parent_rollup_unchanged(self):
+        child = _make_mock_child(cost=0.25)
+        parent, _ = self._run(child)
+        self.assertAlmostEqual(parent.session_estimated_cost_usd, 0.25, places=6)
+        self.assertEqual(parent.session_cost_source, "subagent")
+        self.assertEqual(parent.session_cost_status, "estimated")
+
+    def test_non_numeric_child_cost_degrades_to_zero(self):
+        child = _make_mock_child(cost=0.0)
+        child.session_estimated_cost_usd = "not-a-number"
+        _, result = self._run(child)
+        entry = result["results"][0]
+        self.assertEqual(entry["cost_usd"], 0.0)
+
+
+if __name__ == "__main__":
+    unittest.main()
diff --git a/tools/delegate_tool.py b/tools/delegate_tool.py
index 8d612e8bcbc30..a3e84b5675eef 100644
--- a/tools/delegate_tool.py
+++ b/tools/delegate_tool.py
@@ -2590,6 +2590,17 @@ def _run_single_child(
                 else 0.0
             ),
         }
+        # Per-delegation spend, serialized back to the model alongside
+        # tokens/api_calls so the parent can see what each delegation cost.
+        # Mirrors _child_cost_usd (which is stripped pre-serialization and
+        # only feeds the parent session rollup).
+        # Inspired by: Perplexity Agent API result shape (idea-level).
+        entry["cost_usd"] = round(entry["_child_cost_usd"], 6)
+        _cost_status = getattr(child, "session_cost_status", None)
+        entry["cost_status"] = (
+            _cost_status if isinstance(_cost_status, str) and _cost_status
+            else "unknown"
+        )
         if status == "failed":
             entry["error"] = result.get("error", "Subagent did not produce a response.")