From b5a7d2ab34740619b1f5de52663e2617015bcf36 Mon Sep 17 00:00:00 2001
From: yuneng-jiang <yuneng.jiang@gmail.com>
Date: Wed, 21 Jan 2026 10:44:18 -0800
Subject: [PATCH 1/3] Paginating model/info endpoint

---
 litellm/proxy/proxy_server.py                 |  31 ++-
 tests/test_litellm/proxy/test_proxy_server.py | 225 ++++++++++++++++++
 2 files changed, 254 insertions(+), 2 deletions(-)

diff --git a/litellm/proxy/proxy_server.py b/litellm/proxy/proxy_server.py
index eef2af89799..c4f2b45f388 100644
--- a/litellm/proxy/proxy_server.py
+++ b/litellm/proxy/proxy_server.py
@@ -7522,6 +7522,8 @@ async def model_info_v2(
         False, description="Return all models across all teams user is in."
     ),
     debug: Optional[bool] = False,
+    page: int = Query(1, description="Page number", ge=1),
+    size: int = Query(50, description="Page size", ge=1),
 ):
     """
     BETA ENDPOINT. Might change unexpectedly. Use `/v1/model/info` for now.
@@ -7530,7 +7532,13 @@ async def model_info_v2(
 
     # Return empty data array when no models are configured (graceful handling for fresh installs)
     if llm_router is None or not llm_router.model_list:
-        return {"data": []}
+        return {
+            "data": [],
+            "total_count": 0,
+            "current_page": page,
+            "total_pages": 0,
+            "size": size,
+        }
 
     if prisma_client is None:
         raise HTTPException(
@@ -7620,7 +7628,26 @@ async def model_info_v2(
         )
 
     verbose_proxy_logger.debug("all_models: %s", all_models)
-    return {"data": all_models}
+    
+    total_count = len(all_models)
+    
+    skip = (page - 1) * size
+    
+    total_pages = -(-total_count // size) if total_count > 0 else 0
+    
+    paginated_models = all_models[skip : skip + size]
+    
+    verbose_proxy_logger.debug(
+        f"Pagination: skip={skip}, take={size}, total_count={total_count}, total_pages={total_pages}"
+    )
+    
+    return {
+        "data": paginated_models,
+        "total_count": total_count,
+        "current_page": page,
+        "total_pages": total_pages,
+        "size": size,
+    }
 
 
 @router.get(
diff --git a/tests/test_litellm/proxy/test_proxy_server.py b/tests/test_litellm/proxy/test_proxy_server.py
index 751a9033871..d8970a76a9a 100644
--- a/tests/test_litellm/proxy/test_proxy_server.py
+++ b/tests/test_litellm/proxy/test_proxy_server.py
@@ -3216,3 +3216,228 @@ async def test_get_hierarchical_router_settings():
         prisma_client=mock_prisma_client,
     )
     assert result is None
+
+
+@pytest.mark.asyncio
+async def test_model_info_v2_pagination_basic(monkeypatch):
+    """
+    Test basic pagination functionality for /v2/model/info endpoint.
+    Tests multiple pages with different page sizes.
+    """
+    from unittest.mock import AsyncMock, MagicMock
+
+    from litellm.proxy._types import UserAPIKeyAuth
+    from litellm.proxy.proxy_server import app, proxy_config, user_api_key_auth
+
+    # Create 75 mock models for testing pagination
+    mock_models = [
+        {
+            "model_name": f"model-{i}",
+            "litellm_params": {"model": f"gpt-{i}"},
+            "model_info": {"id": f"model-{i}"},
+        }
+        for i in range(1, 76)  # 75 models total
+    ]
+
+    # Mock llm_router
+    mock_router = MagicMock()
+    mock_router.model_list = mock_models
+
+    # Mock prisma_client
+    mock_prisma_client = MagicMock()
+
+    # Mock proxy_config.get_config
+    mock_get_config = AsyncMock(return_value={})
+
+    # Mock user authentication
+    mock_user_api_key_dict = MagicMock(spec=UserAPIKeyAuth)
+    mock_user_api_key_dict.user_id = "test-user"
+    mock_user_api_key_dict.api_key = "test-key"
+    mock_user_api_key_dict.team_models = []
+    mock_user_api_key_dict.models = []
+
+    # Apply monkeypatches
+    monkeypatch.setattr("litellm.proxy.proxy_server.llm_router", mock_router)
+    monkeypatch.setattr("litellm.proxy.proxy_server.prisma_client", mock_prisma_client)
+    monkeypatch.setattr("litellm.proxy.proxy_server.user_model", None)
+    monkeypatch.setattr(proxy_config, "get_config", mock_get_config)
+
+    # Override auth dependency
+    original_overrides = app.dependency_overrides.copy()
+    app.dependency_overrides[user_api_key_auth] = lambda: mock_user_api_key_dict
+
+    client = TestClient(app)
+    try:
+        # Test page 1 with size 25 (should return models 1-25)
+        response = client.get("/v2/model/info", params={"page": 1, "size": 25})
+        assert response.status_code == 200
+        data = response.json()
+        assert data["total_count"] == 75
+        assert data["current_page"] == 1
+        assert data["size"] == 25
+        assert data["total_pages"] == 3  # ceil(75/25) = 3
+        assert len(data["data"]) == 25
+        assert data["data"][0]["model_name"] == "model-1"
+        assert data["data"][24]["model_name"] == "model-25"
+
+        # Test page 2 with size 25 (should return models 26-50)
+        response = client.get("/v2/model/info", params={"page": 2, "size": 25})
+        assert response.status_code == 200
+        data = response.json()
+        assert data["total_count"] == 75
+        assert data["current_page"] == 2
+        assert data["size"] == 25
+        assert data["total_pages"] == 3
+        assert len(data["data"]) == 25
+        assert data["data"][0]["model_name"] == "model-26"
+        assert data["data"][24]["model_name"] == "model-50"
+
+        # Test page 3 with size 25 (should return models 51-75)
+        response = client.get("/v2/model/info", params={"page": 3, "size": 25})
+        assert response.status_code == 200
+        data = response.json()
+        assert data["total_count"] == 75
+        assert data["current_page"] == 3
+        assert data["size"] == 25
+        assert data["total_pages"] == 3
+        assert len(data["data"]) == 25
+        assert data["data"][0]["model_name"] == "model-51"
+        assert data["data"][24]["model_name"] == "model-75"
+
+        # Test different page size (size 10)
+        response = client.get("/v2/model/info", params={"page": 1, "size": 10})
+        assert response.status_code == 200
+        data = response.json()
+        assert data["total_count"] == 75
+        assert data["current_page"] == 1
+        assert data["size"] == 10
+        assert data["total_pages"] == 8  # ceil(75/10) = 8
+        assert len(data["data"]) == 10
+
+    finally:
+        app.dependency_overrides = original_overrides
+
+
+@pytest.mark.asyncio
+async def test_model_info_v2_pagination_edge_cases(monkeypatch):
+    """
+    Test edge cases for pagination in /v2/model/info endpoint.
+    Tests empty results, last page with partial results, and boundary conditions.
+    """
+    from unittest.mock import AsyncMock, MagicMock
+
+    from litellm.proxy._types import UserAPIKeyAuth
+    from litellm.proxy.proxy_server import app, proxy_config, user_api_key_auth
+
+    # Mock prisma_client
+    mock_prisma_client = MagicMock()
+
+    # Mock user authentication
+    mock_user_api_key_dict = MagicMock(spec=UserAPIKeyAuth)
+    mock_user_api_key_dict.user_id = "test-user"
+    mock_user_api_key_dict.api_key = "test-key"
+    mock_user_api_key_dict.team_models = []
+    mock_user_api_key_dict.models = []
+
+    # Mock proxy_config.get_config
+    mock_get_config = AsyncMock(return_value={})
+
+    # Apply monkeypatches
+    monkeypatch.setattr("litellm.proxy.proxy_server.prisma_client", mock_prisma_client)
+    monkeypatch.setattr("litellm.proxy.proxy_server.user_model", None)
+    monkeypatch.setattr(proxy_config, "get_config", mock_get_config)
+
+    # Override auth dependency
+    original_overrides = app.dependency_overrides.copy()
+    app.dependency_overrides[user_api_key_auth] = lambda: mock_user_api_key_dict
+
+    client = TestClient(app)
+    try:
+        # Test Case 1: Empty model list (no models configured)
+        mock_router_empty = MagicMock()
+        mock_router_empty.model_list = []
+        monkeypatch.setattr("litellm.proxy.proxy_server.llm_router", mock_router_empty)
+
+        response = client.get("/v2/model/info", params={"page": 1, "size": 25})
+        assert response.status_code == 200
+        data = response.json()
+        assert data["total_count"] == 0
+        assert data["current_page"] == 1
+        assert data["size"] == 25
+        assert data["total_pages"] == 0
+        assert len(data["data"]) == 0
+
+        # Test Case 2: Last page with partial results (23 models, page size 10)
+        mock_models_partial = [
+            {
+                "model_name": f"model-{i}",
+                "litellm_params": {"model": f"gpt-{i}"},
+                "model_info": {"id": f"model-{i}"},
+            }
+            for i in range(1, 24)  # 23 models total
+        ]
+        mock_router_partial = MagicMock()
+        mock_router_partial.model_list = mock_models_partial
+        monkeypatch.setattr("litellm.proxy.proxy_server.llm_router", mock_router_partial)
+
+        # Page 1 should have 10 models
+        response = client.get("/v2/model/info", params={"page": 1, "size": 10})
+        assert response.status_code == 200
+        data = response.json()
+        assert data["total_count"] == 23
+        assert data["current_page"] == 1
+        assert data["total_pages"] == 3  # ceil(23/10) = 3
+        assert len(data["data"]) == 10
+
+        # Page 2 should have 10 models
+        response = client.get("/v2/model/info", params={"page": 2, "size": 10})
+        assert response.status_code == 200
+        data = response.json()
+        assert data["total_count"] == 23
+        assert data["current_page"] == 2
+        assert data["total_pages"] == 3
+        assert len(data["data"]) == 10
+
+        # Page 3 (last page) should have only 3 models
+        response = client.get("/v2/model/info", params={"page": 3, "size": 10})
+        assert response.status_code == 200
+        data = response.json()
+        assert data["total_count"] == 23
+        assert data["current_page"] == 3
+        assert data["total_pages"] == 3
+        assert len(data["data"]) == 3
+        assert data["data"][0]["model_name"] == "model-21"
+        assert data["data"][2]["model_name"] == "model-23"
+
+        # Test Case 3: Page beyond available pages (should return empty data)
+        response = client.get("/v2/model/info", params={"page": 4, "size": 10})
+        assert response.status_code == 200
+        data = response.json()
+        assert data["total_count"] == 23
+        assert data["current_page"] == 4
+        assert data["total_pages"] == 3
+        assert len(data["data"]) == 0  # No data for page beyond total_pages
+
+        # Test Case 4: Single model with page size 1
+        mock_models_single = [
+            {
+                "model_name": "single-model",
+                "litellm_params": {"model": "gpt-4"},
+                "model_info": {"id": "single-model"},
+            }
+        ]
+        mock_router_single = MagicMock()
+        mock_router_single.model_list = mock_models_single
+        monkeypatch.setattr("litellm.proxy.proxy_server.llm_router", mock_router_single)
+
+        response = client.get("/v2/model/info", params={"page": 1, "size": 1})
+        assert response.status_code == 200
+        data = response.json()
+        assert data["total_count"] == 1
+        assert data["current_page"] == 1
+        assert data["total_pages"] == 1
+        assert len(data["data"]) == 1
+        assert data["data"][0]["model_name"] == "single-model"
+
+    finally:
+        app.dependency_overrides = original_overrides

From d0e35751a11db388b49a1c8f85ad9b6bb2f123f6 Mon Sep 17 00:00:00 2001
From: yuneng-jiang <yuneng.jiang@gmail.com>
Date: Wed, 21 Jan 2026 11:02:39 -0800
Subject: [PATCH 2/3] Fixing tests and linting

---
 litellm/proxy/proxy_server.py                 | 126 +++++++-----
 .../proxy/test_empty_model_list.py            |  58 +++++-
 tests/test_litellm/proxy/test_proxy_server.py | 184 ++++++++++++++++++
 3 files changed, 312 insertions(+), 56 deletions(-)

diff --git a/litellm/proxy/proxy_server.py b/litellm/proxy/proxy_server.py
index c4f2b45f388..d168290e8a1 100644
--- a/litellm/proxy/proxy_server.py
+++ b/litellm/proxy/proxy_server.py
@@ -7503,6 +7503,77 @@ async def get_all_team_and_direct_access_models(
     return all_models
 
 
+def _enrich_model_info_with_litellm_data(
+    model: Dict[str, Any], debug: bool = False, llm_router: Optional[Router] = None
+) -> Dict[str, Any]:
+    """
+    Enrich a model dictionary with litellm model info (pricing, context window, etc.)
+    and remove sensitive information.
+    
+    Args:
+        model: Model dictionary to enrich
+        debug: Whether to include debug information like openai_client
+        llm_router: Optional router instance for debug info
+        
+    Returns:
+        Enriched model dictionary with sensitive info removed
+    """
+    # provided model_info in config.yaml
+    model_info = model.get("model_info", {})
+    if debug is True:
+        _openai_client = "None"
+        if llm_router is not None:
+            _openai_client = (
+                llm_router._get_client(
+                    deployment=model, kwargs={}, client_type="async"
+                )
+                or "None"
+            )
+        else:
+            _openai_client = "llm_router_is_None"
+        openai_client = str(_openai_client)
+        model["openai_client"] = openai_client
+
+    # read litellm model_prices_and_context_window.json to get the following:
+    # input_cost_per_token, output_cost_per_token, max_tokens
+    litellm_model_info = get_litellm_model_info(model=model)
+
+    # 2nd pass on the model, try seeing if we can find model in litellm model_cost map
+    if litellm_model_info == {}:
+        # use litellm_param model_name to get model_info
+        litellm_params = model.get("litellm_params", {})
+        litellm_model = litellm_params.get("model", None)
+        try:
+            litellm_model_info = litellm.get_model_info(model=litellm_model)
+        except Exception:
+            litellm_model_info = {}
+    # 3rd pass on the model, try seeing if we can find model but without the "/" in model cost map
+    if litellm_model_info == {}:
+        # use litellm_param model_name to get model_info
+        litellm_params = model.get("litellm_params", {})
+        litellm_model = litellm_params.get("model", None)
+        if litellm_model:
+            split_model = litellm_model.split("/")
+            if len(split_model) > 0:
+                litellm_model = split_model[-1]
+            try:
+                litellm_model_info = litellm.get_model_info(
+                    model=litellm_model, custom_llm_provider=split_model[0]
+                )
+            except Exception:
+                litellm_model_info = {}
+    for k, v in litellm_model_info.items():
+        if k not in model_info:
+            model_info[k] = v
+    model["model_info"] = model_info
+    # don't return the api key / vertex credentials
+    # don't return the llm credentials
+    model = remove_sensitive_info_from_deployment(
+        model, excluded_keys={"litellm_credential_name"}
+    )
+    return model
+
+
 @router.get(
     "/v2/model/info",
     description="v2 - returns models available to the user based on their API key permissions. Shows model info from config.yaml (except api key and api base). Filter to just user-added models with ?user_models_only=true",
@@ -7573,58 +7644,9 @@ async def model_info_v2(
             all_models=all_models,
         )
     # fill in model info based on config.yaml and litellm model_prices_and_context_window.json
-    for _model in all_models:
-        # provided model_info in config.yaml
-        model_info = _model.get("model_info", {})
-        if debug is True:
-            _openai_client = "None"
-            if llm_router is not None:
-                _openai_client = (
-                    llm_router._get_client(
-                        deployment=_model, kwargs={}, client_type="async"
-                    )
-                    or "None"
-                )
-            else:
-                _openai_client = "llm_router_is_None"
-            openai_client = str(_openai_client)
-            _model["openai_client"] = openai_client
-
-        # read litellm model_prices_and_context_window.json to get the following:
-        # input_cost_per_token, output_cost_per_token, max_tokens
-        litellm_model_info = get_litellm_model_info(model=_model)
-
-        # 2nd pass on the model, try seeing if we can find model in litellm model_cost map
-        if litellm_model_info == {}:
-            # use litellm_param model_name to get model_info
-            litellm_params = _model.get("litellm_params", {})
-            litellm_model = litellm_params.get("model", None)
-            try:
-                litellm_model_info = litellm.get_model_info(model=litellm_model)
-            except Exception:
-                litellm_model_info = {}
-        # 3rd pass on the model, try seeing if we can find model but without the "/" in model cost map
-        if litellm_model_info == {}:
-            # use litellm_param model_name to get model_info
-            litellm_params = _model.get("litellm_params", {})
-            litellm_model = litellm_params.get("model", None)
-            split_model = litellm_model.split("/")
-            if len(split_model) > 0:
-                litellm_model = split_model[-1]
-            try:
-                litellm_model_info = litellm.get_model_info(
-                    model=litellm_model, custom_llm_provider=split_model[0]
-                )
-            except Exception:
-                litellm_model_info = {}
-        for k, v in litellm_model_info.items():
-            if k not in model_info:
-                model_info[k] = v
-        _model["model_info"] = model_info
-        # don't return the api key / vertex credentials
-        # don't return the llm credentials
-        _model = remove_sensitive_info_from_deployment(
-            _model, excluded_keys={"litellm_credential_name"}
+    for i, _model in enumerate(all_models):
+        all_models[i] = _enrich_model_info_with_litellm_data(
+            model=_model, debug=debug, llm_router=llm_router
         )
 
     verbose_proxy_logger.debug("all_models: %s", all_models)
diff --git a/tests/test_litellm/proxy/test_empty_model_list.py b/tests/test_litellm/proxy/test_empty_model_list.py
index 6b3e59d3194..dd900d3eb53 100644
--- a/tests/test_litellm/proxy/test_empty_model_list.py
+++ b/tests/test_litellm/proxy/test_empty_model_list.py
@@ -32,7 +32,7 @@ class TestEmptyModelListHandling:
         self, client, monkeypatch
     ):
         """
-        Test that /v2/model/info returns {"data": []} instead of 500
+        Test that /v2/model/info returns paginated empty response instead of 500
         when llm_router is None.
         """
         monkeypatch.setattr("litellm.proxy.proxy_server.llm_router", None)
@@ -56,13 +56,18 @@ class TestEmptyModelListHandling:
             )
 
         assert response.status_code == 200
-        assert response.json() == {"data": []}
+        data = response.json()
+        assert data["data"] == []
+        assert data["total_count"] == 0
+        assert data["current_page"] == 1
+        assert data["total_pages"] == 0
+        assert data["size"] == 50  # default page size
 
     def test_v2_model_info_returns_empty_data_when_model_list_empty(
         self, client, monkeypatch
     ):
         """
-        Test that /v2/model/info returns {"data": []} instead of 500
+        Test that /v2/model/info returns paginated empty response instead of 500
         when llm_router exists but model_list is empty.
         """
         mock_router = MagicMock()
@@ -89,7 +94,52 @@ class TestEmptyModelListHandling:
             )
 
         assert response.status_code == 200
-        assert response.json() == {"data": []}
+        data = response.json()
+        assert data["data"] == []
+        assert data["total_count"] == 0
+        assert data["current_page"] == 1
+        assert data["total_pages"] == 0
+        assert data["size"] == 50  # default page size
+
+    def test_v2_model_info_pagination_with_empty_results(
+        self, client, monkeypatch
+    ):
+        """
+        Test that /v2/model/info pagination parameters work correctly
+        when there are no models (empty results).
+        """
+        mock_router = MagicMock()
+        mock_router.model_list = []
+
+        monkeypatch.setattr("litellm.proxy.proxy_server.llm_router", mock_router)
+        monkeypatch.setattr("litellm.proxy.proxy_server.llm_model_list", [])
+        monkeypatch.setattr("litellm.proxy.proxy_server.user_model", None)
+        monkeypatch.setattr("litellm.proxy.proxy_server.general_settings", {})
+
+        with patch(
+            "litellm.proxy.auth.user_api_key_auth.user_api_key_auth",
+            return_value=MagicMock(
+                user_id="test-user",
+                team_id=None,
+                team_models=[],
+                models=[],
+                user_role="proxy_admin",
+            ),
+        ):
+            # Test with custom pagination parameters
+            response = client.get(
+                "/v2/model/info",
+                params={"page": 2, "size": 25},
+                headers={"Authorization": "Bearer sk-test"},
+            )
+
+        assert response.status_code == 200
+        data = response.json()
+        assert data["data"] == []
+        assert data["total_count"] == 0
+        assert data["current_page"] == 2  # Should respect the page parameter
+        assert data["total_pages"] == 0
+        assert data["size"] == 25  # Should respect the size parameter
 
     def test_model_group_info_returns_empty_data_when_model_list_none(
         self, client, monkeypatch
diff --git a/tests/test_litellm/proxy/test_proxy_server.py b/tests/test_litellm/proxy/test_proxy_server.py
index d8970a76a9a..f1854380efe 100644
--- a/tests/test_litellm/proxy/test_proxy_server.py
+++ b/tests/test_litellm/proxy/test_proxy_server.py
@@ -3441,3 +3441,187 @@ async def test_model_info_v2_pagination_edge_cases(monkeypatch):
 
     finally:
         app.dependency_overrides = original_overrides
+
+
+def test_enrich_model_info_with_litellm_data():
+    """
+    Test the _enrich_model_info_with_litellm_data helper function.
+    Tests model info enrichment, debug mode, and sensitive info removal.
+    """
+    from unittest.mock import MagicMock, patch
+
+    from litellm.proxy.proxy_server import _enrich_model_info_with_litellm_data
+
+    # Test Case 1: Basic model enrichment without debug
+    model = {
+        "model_name": "test-model",
+        "litellm_params": {"model": "gpt-3.5-turbo"},
+        "model_info": {"id": "test-model"},
+        "api_key": "sk-secret-key",  # Should be removed
+    }
+
+    with patch("litellm.proxy.proxy_server.get_litellm_model_info") as mock_get_info, patch(
+        "litellm.proxy.proxy_server.remove_sensitive_info_from_deployment"
+    ) as mock_remove_sensitive:
+        mock_get_info.return_value = {
+            "input_cost_per_token": 0.001,
+            "output_cost_per_token": 0.002,
+            "max_tokens": 4096,
+        }
+        mock_remove_sensitive.return_value = {
+            "model_name": "test-model",
+            "litellm_params": {"model": "gpt-3.5-turbo"},
+            "model_info": {
+                "id": "test-model",
+                "input_cost_per_token": 0.001,
+                "output_cost_per_token": 0.002,
+                "max_tokens": 4096,
+            },
+        }
+
+        result = _enrich_model_info_with_litellm_data(model=model, debug=False)
+
+        # Verify get_litellm_model_info was called
+        mock_get_info.assert_called_once_with(model=model)
+        # Verify remove_sensitive_info_from_deployment was called
+        mock_remove_sensitive.assert_called_once()
+        # Verify result doesn't have api_key
+        assert "api_key" not in result
+        # Verify model_info was enriched
+        assert "input_cost_per_token" in result["model_info"]
+
+    # Test Case 2: Model enrichment with debug mode
+    model_with_debug = {
+        "model_name": "test-model-debug",
+        "litellm_params": {"model": "gpt-4"},
+        "model_info": {},
+    }
+
+    mock_router = MagicMock()
+    mock_client = MagicMock()
+    mock_router._get_client.return_value = mock_client
+
+    with patch("litellm.proxy.proxy_server.get_litellm_model_info") as mock_get_info, patch(
+        "litellm.proxy.proxy_server.remove_sensitive_info_from_deployment"
+    ) as mock_remove_sensitive:
+        mock_get_info.return_value = {}
+        mock_remove_sensitive.return_value = {
+            "model_name": "test-model-debug",
+            "litellm_params": {"model": "gpt-4"},
+            "model_info": {},
+            "openai_client": str(mock_client),
+        }
+
+        result = _enrich_model_info_with_litellm_data(
+            model=model_with_debug, debug=True, llm_router=mock_router
+        )
+
+        # Verify debug info was added
+        mock_remove_sensitive.assert_called_once()
+        call_args = mock_remove_sensitive.call_args[0][0]
+        assert "openai_client" in call_args
+        # Verify router._get_client was called for debug
+        mock_router._get_client.assert_called_once()
+
+    # Test Case 3: Model with fallback to litellm.get_model_info
+    model_fallback = {
+        "model_name": "test-model-fallback",
+        "litellm_params": {"model": "claude-3-opus"},
+        "model_info": {},
+    }
+
+    with patch("litellm.proxy.proxy_server.get_litellm_model_info") as mock_get_info, patch(
+        "litellm.get_model_info"
+    ) as mock_litellm_info, patch(
+        "litellm.proxy.proxy_server.remove_sensitive_info_from_deployment"
+    ) as mock_remove_sensitive:
+        # First call returns empty, triggering fallback
+        mock_get_info.return_value = {}
+        mock_litellm_info.return_value = {
+            "input_cost_per_token": 0.015,
+            "output_cost_per_token": 0.075,
+            "max_tokens": 200000,
+        }
+        mock_remove_sensitive.return_value = {
+            "model_name": "test-model-fallback",
+            "litellm_params": {"model": "claude-3-opus"},
+            "model_info": {
+                "input_cost_per_token": 0.015,
+                "output_cost_per_token": 0.075,
+                "max_tokens": 200000,
+            },
+        }
+
+        result = _enrich_model_info_with_litellm_data(model=model_fallback, debug=False)
+
+        # Verify fallback was attempted
+        mock_litellm_info.assert_called_once_with(model="claude-3-opus")
+        # Verify model_info was enriched with fallback data
+        call_args = mock_remove_sensitive.call_args[0][0]
+        assert call_args["model_info"]["input_cost_per_token"] == 0.015
+
+    # Test Case 4: Model with split model name fallback
+    model_split = {
+        "model_name": "test-model-split",
+        "litellm_params": {"model": "azure/gpt-4"},
+        "model_info": {},
+    }
+
+    with patch("litellm.proxy.proxy_server.get_litellm_model_info") as mock_get_info, patch(
+        "litellm.get_model_info"
+    ) as mock_litellm_info, patch(
+        "litellm.proxy.proxy_server.remove_sensitive_info_from_deployment"
+    ) as mock_remove_sensitive:
+        # Both first and second pass return empty, triggering third pass
+        mock_get_info.return_value = {}
+        # Second pass (no split)
+        mock_litellm_info.side_effect = [
+            {},  # First call returns empty
+            {"max_tokens": 8192},  # Third pass with split succeeds
+        ]
+        mock_remove_sensitive.return_value = {
+            "model_name": "test-model-split",
+            "litellm_params": {"model": "azure/gpt-4"},
+            "model_info": {"max_tokens": 8192},
+        }
+
+        result = _enrich_model_info_with_litellm_data(model=model_split, debug=False)
+
+        # Verify third pass was attempted with split model name
+        assert mock_litellm_info.call_count == 2
+        # Check that second call used split model name
+        second_call = mock_litellm_info.call_args_list[1]
+        assert second_call[1]["model"] == "gpt-4"
+        assert second_call[1]["custom_llm_provider"] == "azure"
+
+    # Test Case 5: Model with existing model_info (should preserve existing keys)
+    model_existing = {
+        "model_name": "test-model-existing",
+        "litellm_params": {"model": "gpt-3.5-turbo"},
+        "model_info": {"id": "existing-id", "custom_key": "custom_value"},
+    }
+
+    with patch("litellm.proxy.proxy_server.get_litellm_model_info") as mock_get_info, patch(
+        "litellm.proxy.proxy_server.remove_sensitive_info_from_deployment"
+    ) as mock_remove_sensitive:
+        mock_get_info.return_value = {
+            "input_cost_per_token": 0.001,
+            "id": "new-id",  # Should not override existing "id"
+        }
+        mock_remove_sensitive.return_value = {
+            "model_name": "test-model-existing",
+            "litellm_params": {"model": "gpt-3.5-turbo"},
+            "model_info": {
+                "id": "existing-id",  # Existing key preserved
+                "custom_key": "custom_value",  # Existing key preserved
+                "input_cost_per_token": 0.001,  # New key added
+            },
+        }
+
+        result = _enrich_model_info_with_litellm_data(model=model_existing, debug=False)
+
+        # Verify existing keys are preserved
+        call_args = mock_remove_sensitive.call_args[0][0]
+        assert call_args["model_info"]["id"] == "existing-id"
+        assert call_args["model_info"]["custom_key"] == "custom_value"
+        assert call_args["model_info"]["input_cost_per_token"] == 0.001

From 3075b0e5a25462b76824f5236b0416c39c10de38 Mon Sep 17 00:00:00 2001
From: yuneng-jiang <yuneng.jiang@gmail.com>
Date: Wed, 21 Jan 2026 11:22:20 -0800
Subject: [PATCH 3/3] fixing mypy linting

---
 litellm/proxy/proxy_server.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/litellm/proxy/proxy_server.py b/litellm/proxy/proxy_server.py
index d168290e8a1..8ae07121177 100644
--- a/litellm/proxy/proxy_server.py
+++ b/litellm/proxy/proxy_server.py
@@ -7646,7 +7646,7 @@ async def model_info_v2(
     # fill in model info based on config.yaml and litellm model_prices_and_context_window.json
     for i, _model in enumerate(all_models):
         all_models[i] = _enrich_model_info_with_litellm_data(
-            model=_model, debug=debug, llm_router=llm_router
+            model=_model, debug=debug if debug is not None else False, llm_router=llm_router
         )
 
     verbose_proxy_logger.debug("all_models: %s", all_models)