diff --git a/litellm/router_strategy/lowest_cost.py b/litellm/router_strategy/lowest_cost.py
index 8889180a994..a52c6d468ca 100644
--- a/litellm/router_strategy/lowest_cost.py
+++ b/litellm/router_strategy/lowest_cost.py
@@ -262,12 +262,12 @@ class LowestCostLoggingHandler(CustomLogger):
             if item_output_cost is None:
                 item_output_cost = item_litellm_model_cost_map.get("output_cost_per_token", 5.0)
 
-            item_cache_read_cost = None
-            if _deployment.get("litellm_params", {}).get("cache_read_input_token_cost", None):
-                item_cache_read_cost = _deployment.get("litellm_params", {}).get("cache_read_input_token_cost")
-
-            if item_cache_read_cost is None:
-                item_cache_read_cost = item_litellm_model_cost_map.get("cache_read_input_token_cost", item_input_cost)
+            deployment_params = _deployment.get("litellm_params")
+            item_cache_read_cost = (
+                deployment_params.get("cache_read_input_token_cost")
+                if deployment_params and deployment_params.get("cache_read_input_token_cost") is not None
+                else item_litellm_model_cost_map.get("cache_read_input_token_cost", item_input_cost)
+            )
 
             # if litellm["model"] is not in model_cost map -> use item_cost = $10
 
diff --git a/tests/test_litellm/router_strategy/test_lowest_cost.py b/tests/test_litellm/router_strategy/test_lowest_cost.py
index 401466a557d..309780f217d 100644
--- a/tests/test_litellm/router_strategy/test_lowest_cost.py
+++ b/tests/test_litellm/router_strategy/test_lowest_cost.py
@@ -7,19 +7,6 @@ from litellm.caching.caching import DualCache
 from litellm.router_strategy.lowest_cost import LowestCostLoggingHandler
 
 
-def _tied_deployment(deployment_id, cache_read_cost):
-    return {
-        "model_name": "cache-tie-test",
-        "litellm_params": {
-            "model": "openai/tie-model-not-in-cost-map",
-            "input_cost_per_token": 1e-06,
-            "output_cost_per_token": 2e-06,
-            "cache_read_input_token_cost": cache_read_cost,
-        },
-        "model_info": {"id": deployment_id},
-    }
-
-
 @pytest.mark.parametrize("cheaper_cache_first", [True, False])
 @pytest.mark.asyncio
 async def test_cost_routing_breaks_input_output_tie_on_cache_read_cost(cheaper_cache_first):
@@ -27,10 +14,28 @@ async def test_cost_routing_breaks_input_output_tie_on_cache_read_cost(cheaper_c
     Regression test for https://github.com/BerriAI/litellm/issues/38064
 
     Two deployments with identical input+output price must be separated by their
-    cache-read price, not by whichever one happens to be listed first.
+    cache-read price, not by whichever one happens to be listed first. The pricier
+    deployment omits a cache-read price to exercise the input-cost fallback.
     """
-    cheaper = _tied_deployment("cheaper-cache", cache_read_cost=1e-08)
-    pricier = _tied_deployment("pricier-cache", cache_read_cost=1e-07)
+    cheaper = {
+        "model_name": "cache-tie-test",
+        "litellm_params": {
+            "model": "openai/tie-model-not-in-cost-map",
+            "input_cost_per_token": 1e-06,
+            "output_cost_per_token": 2e-06,
+            "cache_read_input_token_cost": 1e-08,
+        },
+        "model_info": {"id": "cheaper-cache"},
+    }
+    pricier = {
+        "model_name": "cache-tie-test",
+        "litellm_params": {
+            "model": "openai/tie-model-not-in-cost-map",
+            "input_cost_per_token": 1e-06,
+            "output_cost_per_token": 2e-06,
+        },
+        "model_info": {"id": "pricier-cache"},
+    }
     model_list = [cheaper, pricier] if cheaper_cache_first else [pricier, cheaper]
 
     logger = LowestCostLoggingHandler(router_cache=DualCache())