diff --git a/litellm/proxy/auth/model_checks.py b/litellm/proxy/auth/model_checks.py
index 1625198892f..69294f021d2 100644
--- a/litellm/proxy/auth/model_checks.py
+++ b/litellm/proxy/auth/model_checks.py
@@ -13,7 +13,7 @@ from litellm.router import Router
 from litellm.router_utils.fallback_event_handlers import get_fallback_model_group
 from litellm.types.router import CredentialLiteLLMParams, LiteLLM_Params
 from litellm.types.utils import LlmProviders
-from litellm.utils import get_valid_models
+from litellm.utils import ProviderConfigManager, get_valid_models
 
 _CREDENTIAL_LITELLM_PARAM_FIELDS = set(CredentialLiteLLMParams.model_fields)
 
@@ -45,7 +45,18 @@ def get_provider_models(provider: str, litellm_params: LiteLLM_Params | None = N
     if provider in litellm.models_by_provider:
         provider_models: Final = get_valid_models(custom_llm_provider=provider, litellm_params=litellm_params)
         return provider_models
-    return None
+
+    # Providers with no static catalog (litellm_proxy, hosted_vllm, ollama, ...) are absent
+    # from models_by_provider by design: their model list only exists behind the provider's
+    # own endpoint. ProviderConfigManager still knows about them, so admit them here instead
+    # of returning None before endpoint discovery is ever attempted.
+    try:
+        llm_provider: Final = LlmProviders(provider)
+    except ValueError:
+        return None
+    if ProviderConfigManager.get_provider_model_info(model=None, provider=llm_provider) is None:
+        return None
+    return get_valid_models(custom_llm_provider=provider, litellm_params=litellm_params)
 
 
 def _get_models_from_access_groups(
diff --git a/tests/test_litellm/proxy/auth/test_model_checks.py b/tests/test_litellm/proxy/auth/test_model_checks.py
index 62073f4bf51..534eeb4fa84 100644
--- a/tests/test_litellm/proxy/auth/test_model_checks.py
+++ b/tests/test_litellm/proxy/auth/test_model_checks.py
@@ -803,3 +803,38 @@ def test_get_complete_model_list_sentinel_only_grants_nothing():
         infer_model_from_keys=False,
     )
     assert result == []
+
+
+def test_get_provider_models_admits_providers_without_a_static_catalog():
+    """litellm_proxy and hosted_vllm have no entry in litellm.models_by_provider
+    (their model list only exists behind the provider's own endpoint), so the
+    static-dict gate must not reject them before endpoint discovery runs.
+
+    Regression check only, not a discovery test: with check_provider_endpoint
+    left at its default (off), get_valid_models never reaches the network and
+    falls back to models_by_provider.get(provider, []) -- an empty list, not
+    None. Before the fix, the gate itself returned None for these providers.
+    """
+    import litellm
+    from litellm.proxy.auth.model_checks import get_provider_models
+    from litellm.types.router import LiteLLM_Params
+
+    assert "litellm_proxy" not in litellm.models_by_provider
+    assert "hosted_vllm" not in litellm.models_by_provider
+
+    result = get_provider_models(
+        "litellm_proxy",
+        litellm_params=LiteLLM_Params(
+            model="litellm_proxy/*",
+            api_base="http://upstream:4000",
+            api_key="sk-upstream",
+        ),
+    )
+
+    assert result == []
+
+
+def test_get_provider_models_returns_none_for_an_unknown_provider():
+    from litellm.proxy.auth.model_checks import get_provider_models
+
+    assert get_provider_models("not-a-real-provider") is None