From d37320f8cf242f310bad3e759c9a7dbbdc27b699 Mon Sep 17 00:00:00 2001 From: ryan-crabbe-berri Date: Fri, 18 Sep 2026 13:08:55 -0700 Subject: [PATCH] fix(proxy): report null cost for unpriced deployments on the /model/info id lookup GET /model/info?litellm_model_id= went through _get_proxy_model_info, a copy of _enrich_model_info_with_litellm_data that missed the unpriced check, so the same deployment read as null in the list and as 0 on the id lookup. The id lookup now delegates to the shared helper, so the two paths cannot drift again --- litellm/proxy/proxy_server.py | 40 +-------------- .../proxy_server/test_routes_model_info.py | 50 +++++++++++++++++++ 2 files changed, 51 insertions(+), 39 deletions(-) diff --git a/litellm/proxy/proxy_server.py b/litellm/proxy/proxy_server.py index f21c907c292..f3b2dbd0b03 100644 --- a/litellm/proxy/proxy_server.py +++ b/litellm/proxy/proxy_server.py @@ -15057,45 +15057,7 @@ def _translate_model_name_for_response(model: dict) -> dict: def _get_proxy_model_info(model: dict) -> dict: - # provided model_info in config.yaml - model_info: Final = model.get("model_info", {}) - - # 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: Final = 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 = {} - discovered_model_info: Final = ( - llm_router.get_discovered_model_info(model_info.get("id")) if llm_router is not None else MappingProxyType({}) - ) - for k, v in MappingProxyType({**litellm_model_info, **discovered_model_info}).items(): - if k not in model_info or (model_info[k] is None and k in discovered_model_info): - model_info[k] = v - model["model_info"] = model_info - # don't return the llm credentials - model = remove_sensitive_info_from_deployment(deployment_dict=model, excluded_keys={"litellm_credential_name"}) - - return _translate_model_name_for_response(model) + return _translate_model_name_for_response(_enrich_model_info_with_litellm_data(model=model, llm_router=llm_router)) def _model_info_json_response(data: Sequence[Mapping[str, object]] | Mapping[str, object]) -> Response: diff --git a/tests/test_litellm/proxy/proxy_server/test_routes_model_info.py b/tests/test_litellm/proxy/proxy_server/test_routes_model_info.py index 203a237113f..75a8657356a 100644 --- a/tests/test_litellm/proxy/proxy_server/test_routes_model_info.py +++ b/tests/test_litellm/proxy/proxy_server/test_routes_model_info.py @@ -323,6 +323,56 @@ def test_model_info_reports_null_cost_for_unpriced_deployment_and_zero_for_decla assert input_cost > 0 and output_cost > 0 +def test_model_info_id_lookup_reports_the_same_cost_as_the_list( + client: TestClient, + auth_as: Callable[[], AbstractContextManager[object]], + monkeypatch: pytest.MonkeyPatch, +) -> None: + """``GET /model/info?litellm_model_id=`` must agree with the ``GET /model/info`` list, so an + unpriced deployment cannot read as null in the list and as free on the id lookup.""" + monkeypatch.setattr(litellm, "model_cost", copy.deepcopy(litellm.model_cost)) + declared: Final = {"input_cost_per_token": 0.001, "output_cost_per_token": 0.002} + free: Final = {"input_cost_per_token": 0, "output_cost_per_token": 0} + router: Final = litellm.Router( + model_list=[ + { + "model_name": name, + "litellm_params": {"model": f"openai/{name}", "api_key": "x", "api_base": "http://vllm", **costs}, + "model_info": {"id": f"{name}-id"}, + } + for name, costs in (("vllm-unpriced", {}), ("vllm-free", free), ("vllm-priced", declared)) + ] + ) + monkeypatch.setattr(proxy_server, "llm_router", router) + monkeypatch.setattr(proxy_server, "llm_model_list", router.get_model_list()) + monkeypatch.setattr(proxy_server, "user_model", None) + + def costs_of(response: httpx.Response) -> dict[str, tuple[object, object]]: + assert response.status_code == 200, response.text + return { + row["model_info"]["id"]: ( + row["model_info"].get("input_cost_per_token"), + row["model_info"].get("output_cost_per_token"), + ) + for row in response.json()["data"] + } + + with auth_as(): + listed: Final = costs_of(client.get("/model/info")) + by_id: Final = { + model_id: costs_of(client.get("/model/info", params={"litellm_model_id": model_id}))[model_id] + for model_id in listed + } + + assert listed == { + "vllm-unpriced-id": (None, None), + "vllm-free-id": (0, 0), + "vllm-priced-id": (declared["input_cost_per_token"], declared["output_cost_per_token"]), + } + assert by_id == listed + _invalidate_model_cost_lowercase_map() + + def test_v1_model_info_star_wildcard_filter_keeps_provider_expansion(monkeypatch): from litellm.proxy._types import SpecialModelNames, UserAPIKeyAuth from litellm.proxy.auth import model_checks