fix(openrouter): pattern-based fix for native OpenRouter model double-stripping

Replace the hardcoded NATIVE_OPENROUTER_MODELS set approach with a
pattern-based check in _get_openai_compatible_provider_info: after
stripping the outer "openrouter/" provider prefix, if the remaining
model name still starts with "openrouter/", return immediately without
further stripping.

This fixes openrouter/openrouter/aurora-alpha, openrouter/openrouter/polaris-alpha,
and any future native OpenRouter models — not just the three hard-coded
ones (auto, free, bodybuilder) from the previous approach.

Fixes #16353

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Author: tombii

litellm/litellm_core_utils/get_llm_provider_logic.py

@@ -504,6 +504,15 @@ def _get_openai_compatible_provider_info(  # noqa: PLR0915
     custom_llm_provider = model.split("/", 1)[0]
     model = model.split("/", 1)[1]
 
+    # If the provider is openrouter and the remaining model name still starts
+    # with "openrouter/", that inner prefix is part of the actual model ID on
+    # the OpenRouter API (e.g. openrouter/openrouter/aurora-alpha →
+    # model="openrouter/aurora-alpha").  Return immediately so the prefix is
+    # not stripped a second time.
+    if custom_llm_provider == "openrouter" and model.startswith("openrouter/"):
+        dynamic_api_key = api_key or get_secret_str("OPENROUTER_API_KEY")
+        return model, custom_llm_provider, dynamic_api_key, api_base
+
     # Check JSON providers FIRST (before hardcoded ones)
     from litellm.llms.openai_like.dynamic_config import create_config_class
     from litellm.llms.openai_like.json_loader import JSONProviderRegistry

tests/test_litellm/llms/openrouter/test_openrouter_provider_routing.py

@@ -0,0 +1,73 @@
+"""
+Tests for OpenRouter model name routing in get_llm_provider.
+
+OpenRouter-native models have IDs that start with "openrouter/" (e.g.
+openrouter/auto, openrouter/free, openrouter/aurora-alpha).  When a user
+configures such a model in LiteLLM they use the double-prefixed form
+"openrouter/openrouter/aurora-alpha".  get_llm_provider must strip only
+the outer "openrouter/" provider prefix and leave the inner one intact,
+so the correct model ID is sent to the OpenRouter API.
+
+See: https://github.com/BerriAI/litellm/issues/16353
+"""
+
+import os
+import sys
+
+import pytest
+
+sys.path.insert(0, os.path.abspath("../../../../.."))
+
+import litellm
+
+
+class TestOpenRouterNativeModelRouting:
+    """get_llm_provider must not double-strip native OpenRouter model names."""
+
+    @pytest.mark.parametrize(
+        "input_model,expected_model",
+        [
+            # Well-known native models
+            ("openrouter/openrouter/auto", "openrouter/auto"),
+            ("openrouter/openrouter/free", "openrouter/free"),
+            ("openrouter/openrouter/bodybuilder", "openrouter/bodybuilder"),
+            # Arbitrary native models — the fix must be pattern-based, not a hardcoded list
+            ("openrouter/openrouter/aurora-alpha", "openrouter/aurora-alpha"),
+            ("openrouter/openrouter/polaris-alpha", "openrouter/polaris-alpha"),
+            ("openrouter/openrouter/some-future-model", "openrouter/some-future-model"),
+        ],
+    )
+    def test_double_prefixed_strips_once(self, input_model, expected_model):
+        """openrouter/openrouter/<model> should yield model=openrouter/<model>."""
+        result_model, provider, _, _ = litellm.get_llm_provider(model=input_model)
+        assert provider == "openrouter"
+        assert result_model == expected_model
+
+    @pytest.mark.parametrize(
+        "input_model,expected_model",
+        [
+            ("openrouter/openrouter/aurora-alpha", "openrouter/aurora-alpha"),
+            ("openrouter/openrouter/auto", "openrouter/auto"),
+        ],
+    )
+    def test_no_double_strip_on_second_call(self, input_model, expected_model):
+        """Simulates two consecutive get_llm_provider calls (bridge → completion)."""
+        model_first, provider, _, _ = litellm.get_llm_provider(model=input_model)
+        assert model_first == expected_model
+
+        model_second, provider2, _, _ = litellm.get_llm_provider(model=model_first)
+        assert provider2 == "openrouter"
+        assert model_second == expected_model
+
+    @pytest.mark.parametrize(
+        "input_model,expected_model",
+        [
+            ("openrouter/anthropic/claude-3-haiku", "anthropic/claude-3-haiku"),
+            ("openrouter/meta-llama/llama-3-70b-instruct", "meta-llama/llama-3-70b-instruct"),
+        ],
+    )
+    def test_regular_models_still_strip_normally(self, input_model, expected_model):
+        """Non-native OpenRouter models should still have their prefix stripped."""
+        result_model, provider, _, _ = litellm.get_llm_provider(model=input_model)
+        assert provider == "openrouter"
+        assert result_model == expected_model