diff --git a/tests/test_litellm/llms/stepfun/test_stepfun_pricing.py b/tests/test_litellm/llms/stepfun/test_stepfun_pricing.py
index 936e2d38d22..da0e36a168e 100644
--- a/tests/test_litellm/llms/stepfun/test_stepfun_pricing.py
+++ b/tests/test_litellm/llms/stepfun/test_stepfun_pricing.py
@@ -1,19 +1,21 @@
 import math
 
+import pytest
+
 import litellm
 from litellm.cost_calculator import cost_per_token
 
 
-def _load_local_model_cost():
-    import os
-
-    os.environ["LITELLM_LOCAL_MODEL_COST_MAP"] = "True"
-    litellm.model_cost = litellm.get_model_cost_map(url="")
+@pytest.fixture
+def local_model_cost_map(monkeypatch):
+    monkeypatch.setenv("LITELLM_LOCAL_MODEL_COST_MAP", "True")
+    monkeypatch.setattr(litellm, "model_cost", litellm.get_model_cost_map(url=""))
+    litellm.get_model_info.cache_clear()
+    yield
+    litellm.get_model_info.cache_clear()
 
 
-def test_stepfun_models_in_model_cost():
-    _load_local_model_cost()
-
+def test_stepfun_models_in_model_cost(local_model_cost_map):
     stepfun_models = [
         "stepfun/step-3.7-flash",
         "stepfun/step-3.5-flash",
@@ -25,9 +27,7 @@ def test_stepfun_models_in_model_cost():
         assert litellm.model_cost[model]["litellm_provider"] == "stepfun"
 
 
-def test_stepfun_step_35_flash_cost_calculation():
-    _load_local_model_cost()
-
+def test_stepfun_step_35_flash_cost_calculation(local_model_cost_map):
     prompt_cost, completion_cost = cost_per_token(
         model="stepfun/step-3.5-flash",
         prompt_tokens=1000000,
@@ -39,9 +39,7 @@ def test_stepfun_step_35_flash_cost_calculation():
     assert math.isclose(completion_cost, 0.30, rel_tol=1e-6)
 
 
-def test_stepfun_step_37_flash_cost_calculation():
-    _load_local_model_cost()
-
+def test_stepfun_step_37_flash_cost_calculation(local_model_cost_map):
     prompt_cost, completion_cost = cost_per_token(
         model="stepfun/step-3.7-flash",
         prompt_tokens=1000000,
@@ -53,9 +51,7 @@ def test_stepfun_step_37_flash_cost_calculation():
     assert math.isclose(completion_cost, 1.15, rel_tol=1e-6)
 
 
-def test_stepfun_step_35_flash_cache_hit_cost():
-    _load_local_model_cost()
-
+def test_stepfun_step_35_flash_cache_hit_cost(local_model_cost_map):
     prompt_cost, _ = cost_per_token(
         model="stepfun/step-3.5-flash",
         prompt_tokens=0,
@@ -67,8 +63,6 @@ def test_stepfun_step_35_flash_cache_hit_cost():
     assert math.isclose(prompt_cost, 0.02, rel_tol=1e-6)
 
 
-def test_stepfun_reasoning_models_support_reasoning():
-    _load_local_model_cost()
-
+def test_stepfun_reasoning_models_support_reasoning(local_model_cost_map):
     for model in ("stepfun/step-3.5-flash", "stepfun/step-3.5-flash-2603", "stepfun/step-3.7-flash"):
         assert litellm.model_cost[model]["supports_reasoning"] is True