Fail fast on non-retriable LiteLLM status codes

Author: yangbaechu

src/lighteval/models/endpoints/litellm_model.py

@@ -38,6 +38,7 @@
 
 
 logger = logging.getLogger(__name__)
+NON_RETRIABLE_STATUS_CODES = {401, 403, 404}
 
 if is_package_available("litellm"):
     import litellm
@@ -242,6 +243,8 @@ def __call_api(self, prompt, return_logits, max_new_tokens, num_samples, stop_se
                         logger.warning(f"{error_string}. Returning empty response.")
                         return LitellmModelResponse()
             except Exception as e:
+                if getattr(e, "status_code", None) in NON_RETRIABLE_STATUS_CODES:
+                    raise
                 wait_time = min(
                     64, self.API_RETRY_SLEEP * (self.API_RETRY_MULTIPLIER**attempt)
                 )  # Exponential backoff with max 64s

tests/unit/models/endpoints/test_litellm_model.py

@@ -0,0 +1,116 @@
+# MIT License
+
+# Copyright (c) 2026 The HuggingFace Team
+
+# Permission is hereby granted, free of charge, to any person obtaining a copy
+# of this software and associated documentation files (the "Software"), to deal
+# in the Software without restriction, including without limitation the rights
+# to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+# copies of the Software, and to permit persons to whom the Software is
+# furnished to do so, subject to the following conditions:
+
+# The above copyright notice and this permission notice shall be included in all
+# copies or substantial portions of the Software.
+
+# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+# AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+# OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+# SOFTWARE.
+
+from types import SimpleNamespace
+from unittest.mock import Mock
+
+import pytest
+
+
+pytest.importorskip("litellm")
+
+from lighteval.models.endpoints.litellm_model import LiteLLMClient, LiteLLMModelConfig, litellm
+
+
+def make_response(content: str):
+    return SimpleNamespace(choices=[SimpleNamespace(message=SimpleNamespace(content=content))])
+
+
+def make_client(cache_dir: str, api_max_retry: int = 3) -> LiteLLMClient:
+    return LiteLLMClient(
+        LiteLLMModelConfig(
+            model_name="openai/gpt-4o-mini",
+            provider="openai",
+            api_key="test-key",
+            api_max_retry=api_max_retry,
+            cache_dir=cache_dir,
+        )
+    )
+
+
+class TestLiteLLMClientRetries:
+    def test_rate_limit_error_still_retries(self, monkeypatch, tmp_path):
+        client = make_client(str(tmp_path), api_max_retry=3)
+        calls = []
+        responses = [
+            litellm.RateLimitError(
+                message="rate limited",
+                llm_provider="openai",
+                model="openai/gpt-4o-mini",
+            ),
+            litellm.RateLimitError(
+                message="rate limited",
+                llm_provider="openai",
+                model="openai/gpt-4o-mini",
+            ),
+            make_response("ok"),
+        ]
+
+        def fake_completion(**kwargs):
+            calls.append(kwargs)
+            response = responses.pop(0)
+            if isinstance(response, Exception):
+                raise response
+            return response
+
+        monkeypatch.setattr("lighteval.models.endpoints.litellm_model.litellm.completion", fake_completion)
+        sleep_calls = []
+        monkeypatch.setattr("lighteval.models.endpoints.litellm_model.time.sleep", sleep_calls.append)
+
+        response = client._LiteLLMClient__call_api(
+            prompt=[{"role": "user", "content": "hi"}],
+            return_logits=False,
+            max_new_tokens=10,
+            num_samples=1,
+            stop_sequence=None,
+        )
+
+        assert response.choices[0].message.content == "ok"
+        assert len(calls) == 3
+        assert sleep_calls == [1.0, 2.0]
+
+    def test_non_retriable_status_code_fails_fast(self, monkeypatch, tmp_path):
+        client = make_client(str(tmp_path))
+        calls = []
+
+        class FakeException(Exception):
+            status_code = 401
+
+        def fake_completion(**kwargs):
+            calls.append(kwargs)
+            raise FakeException("unauthorized")
+
+        monkeypatch.setattr("lighteval.models.endpoints.litellm_model.litellm.completion", fake_completion)
+        sleep = Mock()
+        monkeypatch.setattr("lighteval.models.endpoints.litellm_model.time.sleep", sleep)
+
+        with pytest.raises(FakeException):
+            client._LiteLLMClient__call_api(
+                prompt=[{"role": "user", "content": "hi"}],
+                return_logits=False,
+                max_new_tokens=10,
+                num_samples=1,
+                stop_sequence=None,
+            )
+
+        assert len(calls) == 1
+        sleep.assert_not_called()