test(cost): keep provider exemptions explicit on the release

tests/unit/router_utils/test_reasoning_effort_capability.py

@@ -393,25 +393,32 @@ class TestGpt6AstraAdvertisesItsDocumentedLevels:
     @pytest.mark.parametrize(
         "model,custom_llm_provider",
         [
-            ("azure/gpt-6-sol", "azure"),
-            ("azure/gpt-6-luna", "azure"),
-            ("azure/eu/gpt-6-sol", "azure"),
-            ("azure/eu/gpt-6-luna", "azure"),
-            ("azure_ai/gpt-6-sol", "azure_ai"),
-            ("azure_ai/gpt-6-luna", "azure_ai"),
+            ("azure/gpt-6-astra", "azure"),
+            ("azure/us/gpt-6-astra", "azure"),
+            ("azure/eu/gpt-6-astra", "azure"),
+            ("azure_ai/gpt-6-astra", "azure_ai"),
         ],
     )
-    def test_the_azure_entry_advertises_the_same_levels_as_openai(
+    def test_an_azure_hosted_deployment_advertises_the_levels_both_its_routes_accept(
         self, local_model_cost_map, model, custom_llm_provider
     ):
+        """Verified live on 2026-10-01 against a Foundry gpt-6-astra deployment (model
+        gpt-6-astra-2026-09-03): none returns 400 unsupported_value on chat and on Responses, naming
+        low, medium, high and xhigh as the levels it takes; Responses also honors max with reasoning
+        tokens billed while chat refuses it, as OpenAI's chat does. The flag is route-blind, so the
+        row keeps max off and the advertised set stops at the level every route accepts.
+        Earlier evidence (2026-09-05, commit e79f3ec5205) had none accepted on chat; the deployment
+        no longer behaves that way."""
         from litellm.utils import get_model_info_helper
 
-        azure_info = dict(get_model_info_helper(model=model, custom_llm_provider=custom_llm_provider))
-        openai_info = dict(get_model_info_helper(model=model.rsplit("/", 1)[1], custom_llm_provider="openai"))
+        model_info = dict(get_model_info_helper(model=model, custom_llm_provider=custom_llm_provider))
 
-        assert resolve_supported_reasoning_efforts(
-            azure_info, deployment_is_mapped=True
-        ) == resolve_supported_reasoning_efforts(openai_info, deployment_is_mapped=True)
+        assert resolve_supported_reasoning_efforts(model_info, deployment_is_mapped=True) == (
+            "low",
+            "medium",
+            "high",
+            "xhigh",
+        )
 
 
 class TestGpt6SolAndLunaAdvertiseNoneThroughMax:

tests/unit/test_model_prices_and_context_window.py

@@ -16,7 +16,6 @@ from litellm.litellm_core_utils.fallback_generalizations import match_capability
 from litellm.llms.openai.chat.gpt_5_transformation import is_gpt_reasoning_series_name
 from litellm.llms.openai_like.json_loader import JSONProviderRegistry
 from litellm.router_utils.reasoning_effort_capability import resolve_supported_reasoning_efforts
-from litellm.types.utils import LlmProviders
 
 REPO_ROOT = Path(__file__).parents[2]
 PRICES_PATH = REPO_ROOT / "model_prices_and_context_window.json"
@@ -679,12 +678,10 @@ def unregistered_providers(rows: Mapping[str, object]) -> list[str]:
 
 @pytest.mark.parametrize("path", (PRICES_PATH, BACKUP_PRICES_PATH), ids=("main", "backup"))
 def test_every_cost_map_provider_is_registered(path: Path):
-    supported_providers: Final = frozenset(provider.value for provider in LlmProviders)
     rows: Final = {
         name: entry
         for name, entry in json.loads(path.read_text()).items()
-        if isinstance(entry, dict)
-        and entry.get("litellm_provider", "").split("-", 1)[0] in supported_providers
+        if isinstance(entry, dict) and entry.get("litellm_provider") != "openinfer"
     }
     assert rows
     assert unregistered_providers(rows) == [], (

tests/unit/test_utils.py

@@ -8561,7 +8561,7 @@ def test_models_by_provider():
             continue
         elif k == "sample_spec":
             continue
-        elif v["litellm_provider"] not in litellm.provider_list and not JSONProviderRegistry.exists(v["litellm_provider"]):
+        elif v["litellm_provider"] == "openinfer":
             continue
         elif (
             v["litellm_provider"] == "sagemaker"