feat(ui): add the evaluation mode to the Add Model form

litellm/proxy/health_endpoints/_health_endpoints.py

@@ -2078,6 +2078,7 @@ async def test_model_connection(
         "responses",
         "anthropic_messages",
         "ocr",
+        "evaluation",
     ]
     | None = fastapi.Body(
         None,

tests/integration/providers/test_decisions_wire.py

@@ -252,6 +252,23 @@ def test_each_provider_gets_its_own_path_key_and_body_and_is_billed_from_the_cos
         assert math.isclose(_number(row["spend"]), expected_spend, rel_tol=1e-9), row
 
 
+def test_test_connection_evaluation_mode_uses_typesafe_decisions_path(gateway: Gateway) -> None:
+    provider: Final = _PROVIDERS[1]
+    with gateway.scenario() as scenario:
+        handle: Final = _register(scenario, _answer_body(provider))
+        model: Final = _deployment(scenario, handle, provider)
+        response: Final = gateway.request(
+            "POST",
+            "/health/test_connection",
+            {"litellm_params": {"model": model}, "mode": "evaluation"},
+        )
+
+        assert response.status_code == 200, response.text
+        assert response.json()["status"] == "success", response.text
+        (call,) = _upstream_calls(gateway, handle)
+        assert call["path"] == f"/{handle.scenario_id}{provider.path}"
+
+
 def test_repeated_identical_requests_each_reach_the_upstream_and_are_each_billed(gateway: Gateway) -> None:
     with gateway.scenario() as scenario:
         handle: Final = _register(scenario, _answer_body(_PERPLEXITY))

tests/unit/proxy/health_endpoints/test_health_endpoints.py

@@ -2,7 +2,7 @@
 import copy
 import json
 import time
-from collections.abc import Iterator, Mapping, Sequence
+from collections.abc import Awaitable, Iterator, Mapping, Sequence
 from contextlib import contextmanager
 from datetime import datetime, timedelta
 from types import MappingProxyType, SimpleNamespace
@@ -705,6 +705,10 @@ def _test_connection_probe(
 ) -> Iterator[AsyncMock]:
     from litellm.types.router import Deployment, LiteLLM_Params
 
+    async def run_health_check(awaitable: Awaitable[object], _timeout: float) -> dict[str, str]:
+        await awaitable
+        return {"status": "healthy"}
+
     router: Final = MagicMock()
     router.get_deployment.side_effect = lambda model_id: (
         Deployment(
@@ -727,7 +731,7 @@ def _test_connection_probe(
         patch("litellm.proxy.health_endpoints._health_endpoints.litellm.ahealth_check", ahealth_check),
         patch(
             "litellm.proxy.health_endpoints._health_endpoints.run_with_timeout",
-            AsyncMock(return_value={"status": "healthy"}),
+            AsyncMock(side_effect=run_health_check),
         ),
     ):
         yield ahealth_check
@@ -874,6 +878,28 @@ async def test_test_model_connection_request_mode_wins_over_resolved_mode():
     assert ahealth_check.call_args.kwargs["mode"] == "chat"
 
 
+@pytest.mark.asyncio
+async def test_test_model_connection_evaluation_mode_uses_decisions_handler():
+    deployment: Final = MappingProxyType(
+        {
+            "model_name": "typesafe/jev-latest",
+            "litellm_params": {"model": "typesafe/jev-latest", "api_key": "fake-typesafe-key"},
+            "model_info": {"id": "typesafe-jev-id"},
+        }
+    )
+    with _test_connection_probe(deployment) as ahealth_check:
+        result: Final = await health_test_model_connection(
+            request=MagicMock(),
+            mode="evaluation",
+            litellm_params={"model": "typesafe/jev-latest"},
+            model_info={"id": "typesafe-jev-id"},
+            user_api_key_dict=UserAPIKeyAuth(user_id="test-user", token="test-token"),
+        )
+
+    assert result["status"] == "success"
+    assert ahealth_check.call_args.kwargs["mode"] == "evaluation"
+
+
 @pytest.mark.asyncio
 async def test_test_model_connection_uses_loaded_deployment_team_id():
     """
@@ -4665,6 +4691,41 @@ def test_test_model_connection_accepts_image_edit_mode(monkeypatch):
     assert response.json()["status"] == "success"
 
 
+def test_test_model_connection_accepts_evaluation_mode(monkeypatch):
+    monkeypatch.setattr(litellm, "disable_aiohttp_transport", True)
+    litellm.in_memory_llm_clients_cache.flush_cache()
+
+    app = FastAPI()
+    app.include_router(_health_endpoints_module.router)
+    app.dependency_overrides[user_api_key_auth] = lambda: UserAPIKeyAuth(user_role=LitellmUserRoles.PROXY_ADMIN)
+    client = TestClient(app)
+
+    with (
+        patch(  # test-quality-ok: endpoint reads the proxy-global DB client and 500s when it is None; it has no injection seam
+            "litellm.proxy.proxy_server.prisma_client", MagicMock()
+        ),
+        respx.mock(assert_all_called=True) as respx_mock,
+    ):
+        upstream = respx_mock.post("https://api.typesafe.ai/v1/systemone").respond(
+            json={
+                "model": "jev-latest",
+                "answers": {"reachable": {"type": "noul", "noul": 1.0}},
+                "usage": {"input_tokens": 12, "output_tokens": 1},
+            }
+        )
+        response = client.post(
+            "/health/test_connection",
+            json={
+                "mode": "evaluation",
+                "litellm_params": {"model": "typesafe/jev-latest", "api_key": "sk-test"},
+            },
+        )
+
+    assert response.status_code == 200, response.text
+    assert response.json()["status"] == "success"
+    assert upstream.called
+
+
 def _pointfive_admin() -> UserAPIKeyAuth:
     return UserAPIKeyAuth(token="admin-token", user_id="admin-user", user_role=LitellmUserRoles.PROXY_ADMIN)
 

ui/litellm-dashboard/src/components/add_model/AddModelForm.integration.test.tsx

@@ -312,6 +312,18 @@
     expect(await screen.findByRole("button", { name: "Add Model" })).toBeInTheDocument();
   });
 
+  it("offers the Evaluation decisions mode", async () => {
+    const mockUseAuthorized = vi.mocked(await import("@/app/(dashboard)/hooks/useAuthorized"));
+    mockUseAuthorized.default.mockReturnValue(mockAuthorizedUser("proxy_admin", "user-1", true));
+
+    renderWithProviders(<AddModelForm {...createTestProps()} />);
+
+    await screen.findByText("Provider");
+    await userEvent.click(screen.getByRole("combobox", { name: "Mode" }));
+
+    expect(await screen.findByRole("option", { name: "Evaluation - /v1/decisions", exact: true })).toBeInTheDocument();
+  });
+
   it("shows only the Close button in the connection test dialog footer", async () => {
     const mockUseAuthorized = vi.mocked(await import("@/app/(dashboard)/hooks/useAuthorized"));
     mockUseAuthorized.default.mockReturnValue(mockAuthorizedUser("proxy_admin", "user-1", true));
@@ -321,7 +333,7 @@
     await userEvent.click(await screen.findByTestId("test-connect-btn"));
 
     const dialog = await screen.findByRole("dialog");
     const footer = dialog.querySelector('[data-slot="dialog-footer"]');
     expect(footer).not.toBeNull();
     expect(footer!.textContent?.trim()).toBe("Close");
   });

ui/litellm-dashboard/src/components/add_model/add_model_modes.tsx

@@ -13,6 +13,7 @@ export const TEST_MODES = [
   { value: "batch", label: "Batch - /batch" },
   { value: "anthropic_messages", label: "Anthropic Messages - /v1/messages" },
   { value: "ocr", label: "OCR - /ocr" },
+  { value: "evaluation", label: "Evaluation - /v1/decisions" },
 ];
 
 // Define the available auto router routing strategies

ui/litellm-dashboard/src/components/add_model/handle_add_model_submit.test.tsx

@@ -102,6 +102,20 @@ describe("prepareModelAddRequest", () => {
     expect(deployment.litellmParamsObj.timeout).toBe(5);
   });
 
+  it("saves the selected mode under model_info", async () => {
+    const formValues = {
+      model_mappings: [{ public_name: "Jev", litellm_model: "typesafe/jev-latest" }],
+      mode: "evaluation",
+    };
+
+    const deployments = await prepareModelAddRequest({ ...formValues }, "token", null);
+
+    expect(deployments).toHaveLength(1);
+    const [deployment] = deployments!;
+    expect(deployment.modelInfoObj.mode).toBe("evaluation");
+    expect(deployment.litellmParamsObj).not.toHaveProperty("mode");
+  });
+
   it.each([
     ["OpenAI", "openai/*"],
     ["Azure_AI_Studio", "azure_ai/*"],