fix url construction for trailing v1 base urls

litellm/llms/openai/chat/gpt_transformation.py

@@ -721,14 +721,17 @@ class OpenAIGPTConfig(BaseLLMModelInfo, BaseConfig):
         if api_key is None:
             api_key = get_secret_str("OPENAI_API_KEY")
 
-        # Strip api_base to just the base URL (scheme + host + port)
-        parsed_url = httpx.URL(api_base)
-        base_url = f"{parsed_url.scheme}://{parsed_url.host}"
-        if parsed_url.port:
-            base_url += f":{parsed_url.port}"
+        # Remove trailing slash from api_base if present
+        api_base = api_base.rstrip("/")
+
+        # If api_base already ends with /v1, just append /models
+        if api_base.endswith("/v1"):
+            url = f"{api_base}/models"
+        else:
+            url = f"{api_base}/v1/models"
 
         response = litellm.module_level_client.get(
-            url=f"{base_url}/v1/models",
+            url=url,
             headers={"Authorization": f"Bearer {api_key}"},
         )
 

tests/litellm/llms/test_openai_get_models_url.py

@@ -0,0 +1,53 @@
+from unittest.mock import MagicMock, patch
+
+from litellm.llms.openai.chat.gpt_transformation import OpenAIGPTConfig
+
+
+class TestGetModelsUrl:
+    """Test that get_models() constructs the correct /v1/models URL from api_base."""
+
+    def _get_url_for_api_base(self, api_base: str) -> str:
+        """Call get_models() with a mocked HTTP client and return the URL it used."""
+        mock_response = MagicMock()
+        mock_response.status_code = 200
+        mock_response.json.return_value = {"data": []}
+
+        with patch("litellm.module_level_client.get", return_value=mock_response) as mock_get:
+            OpenAIGPTConfig().get_models(api_key="fake-key", api_base=api_base)
+            return mock_get.call_args.kwargs["url"]
+
+    def test_plain_base_url(self):
+        """Standard OpenAI base URL without path should get /v1/models appended."""
+        assert self._get_url_for_api_base("https://api.openai.com") == "https://api.openai.com/v1/models"
+
+    def test_trailing_slash(self):
+        """Trailing slash should be stripped before appending /v1/models."""
+        assert self._get_url_for_api_base("https://api.openai.com/") == "https://api.openai.com/v1/models"
+
+    def test_v1_already_present(self):
+        """api_base ending in /v1 should only get /models appended, not /v1/models."""
+        assert self._get_url_for_api_base("https://api.openai.com/v1") == "https://api.openai.com/v1/models"
+
+    def test_v1_with_trailing_slash(self):
+        """api_base ending in /v1/ should be normalized and get /models appended."""
+        assert self._get_url_for_api_base("https://api.openai.com/v1/") == "https://api.openai.com/v1/models"
+
+    def test_subpath_with_v1(self):
+        """api_base with a sub-path ending in /v1 should preserve the full path."""
+        assert self._get_url_for_api_base("https://opencode.ai/zen/v1") == "https://opencode.ai/zen/v1/models"
+
+    def test_subpath_with_v1_trailing_slash(self):
+        """Sub-path with /v1/ should be normalized and get /models appended."""
+        assert self._get_url_for_api_base("https://opencode.ai/zen/v1/") == "https://opencode.ai/zen/v1/models"
+
+    def test_subpath_without_v1(self):
+        """Sub-path without /v1 should get /v1/models appended."""
+        assert self._get_url_for_api_base("https://opencode.ai/zen") == "https://opencode.ai/zen/v1/models"
+
+    def test_localhost_with_port_and_v1(self):
+        """Localhost with port and /v1 should only get /models appended."""
+        assert self._get_url_for_api_base("http://localhost:11434/v1") == "http://localhost:11434/v1/models"
+
+    def test_localhost_without_v1(self):
+        """Localhost with port but no /v1 should get /v1/models appended."""
+        assert self._get_url_for_api_base("http://localhost:11434") == "http://localhost:11434/v1/models"