diff --git a/gateway/platforms/api_server.py b/gateway/platforms/api_server.py
index 4510361a6..c5d28b0aa 100644
--- a/gateway/platforms/api_server.py
+++ b/gateway/platforms/api_server.py
@@ -1108,6 +1108,18 @@ class APIServerAdapter(BasePlatformAdapter):
         reasoning_config = GatewayRunner._load_reasoning_config()
         model = _resolve_gateway_model()
 
+        # When the primary provider's auth fails (expired token / 429 quota
+        # cap), _resolve_runtime_agent_kwargs() falls through to the fallback
+        # provider chain, whose runtime dict carries its own ``model`` key.
+        # Pop it and let it override the config model, mirroring the native
+        # gateway path (_resolve_session_agent_runtime in run.py). Otherwise
+        # the explicit ``model=model`` below collides with the ``**runtime_kwargs``
+        # spread → "got multiple values for keyword argument 'model'", 500ing
+        # every /v1/chat/completions request while a fallback is active.
+        runtime_model = runtime_kwargs.pop("model", None)
+        if runtime_model:
+            model = runtime_model
+
         user_config = _load_gateway_config()
         enabled_toolsets = sorted(_get_platform_tools(user_config, "api_server"))
 
diff --git a/scripts/release.py b/scripts/release.py
index 391b1f5dc..adc40e35a 100755
--- a/scripts/release.py
+++ b/scripts/release.py
@@ -318,6 +318,7 @@ AUTHOR_MAP = {
     "alelpoan@proton.me": "alelpoan",
     "aman@abacus.ai": "Aman113114-IITD",
     "octavio.turra@gmail.com": "octavioturra",
+    "275877312+ryo-solo@users.noreply.github.com": "ryo-solo",
     "524706+Twanislas@users.noreply.github.com": "Twanislas",
     "9592417+adam91holt@users.noreply.github.com": "adam91holt",
     "kchuang1015@users.noreply.github.com": "kchuang1015",
diff --git a/tests/gateway/test_api_server.py b/tests/gateway/test_api_server.py
index c0a2f52d6..25cbd3ec9 100644
--- a/tests/gateway/test_api_server.py
+++ b/tests/gateway/test_api_server.py
@@ -400,6 +400,84 @@ class TestAdapterInit:
         assert isinstance(agent, FakeAgent)
         assert captured["max_iterations"] == 200
 
+    def test_create_agent_handles_fallback_model_kwarg_collision(self, monkeypatch):
+        """When the primary provider auth-fails, _resolve_runtime_agent_kwargs()
+        returns a runtime dict that carries its own ``model`` key. _create_agent
+        must pop it and let it override the config model — otherwise the explicit
+        ``model=`` collides with ``**runtime_kwargs`` and every request 500s with
+        "got multiple values for keyword argument 'model'"."""
+        captured = {}
+
+        class FakeAgent:
+            def __init__(self, **kwargs):
+                captured.update(kwargs)
+
+        monkeypatch.setattr("run_agent.AIAgent", FakeAgent)
+        monkeypatch.setattr(
+            "gateway.run._resolve_runtime_agent_kwargs",
+            lambda: {
+                "provider": "openrouter",
+                "base_url": "https://openrouter.ai/api/v1",
+                "api_mode": "chat_completions",
+                "model": "anthropic/claude-haiku",  # from the fallback entry
+            },
+        )
+        monkeypatch.setattr("gateway.run._resolve_gateway_model", lambda: "primary/model")
+        monkeypatch.setattr("gateway.run._load_gateway_config", lambda: {})
+        monkeypatch.setattr(
+            "gateway.run.GatewayRunner._load_reasoning_config",
+            staticmethod(lambda: {}),
+        )
+        monkeypatch.setattr("gateway.run.GatewayRunner._load_fallback_model", staticmethod(lambda: None))
+        monkeypatch.setattr("gateway.run._current_max_iterations", lambda: 90)
+        monkeypatch.setattr("hermes_cli.tools_config._get_platform_tools", lambda *_: set())
+
+        adapter = APIServerAdapter(PlatformConfig(enabled=True))
+        monkeypatch.setattr(adapter, "_ensure_session_db", lambda: None)
+
+        # Must not raise TypeError on the duplicate 'model' kwarg.
+        agent = adapter._create_agent(session_id="api-session")
+
+        assert isinstance(agent, FakeAgent)
+        # Fallback model overrides the config model, mirroring the native path.
+        assert captured["model"] == "anthropic/claude-haiku"
+
+    def test_create_agent_keeps_config_model_when_runtime_omits_it(self, monkeypatch):
+        """Happy path (no fallback active): runtime_kwargs has no 'model', so the
+        resolved gateway model is used unchanged. Regression guard for the pop."""
+        captured = {}
+
+        class FakeAgent:
+            def __init__(self, **kwargs):
+                captured.update(kwargs)
+
+        monkeypatch.setattr("run_agent.AIAgent", FakeAgent)
+        monkeypatch.setattr(
+            "gateway.run._resolve_runtime_agent_kwargs",
+            lambda: {
+                "provider": "openrouter",
+                "base_url": "https://openrouter.ai/api/v1",
+                "api_mode": "chat_completions",
+            },
+        )
+        monkeypatch.setattr("gateway.run._resolve_gateway_model", lambda: "primary/model")
+        monkeypatch.setattr("gateway.run._load_gateway_config", lambda: {})
+        monkeypatch.setattr(
+            "gateway.run.GatewayRunner._load_reasoning_config",
+            staticmethod(lambda: {}),
+        )
+        monkeypatch.setattr("gateway.run.GatewayRunner._load_fallback_model", staticmethod(lambda: None))
+        monkeypatch.setattr("gateway.run._current_max_iterations", lambda: 90)
+        monkeypatch.setattr("hermes_cli.tools_config._get_platform_tools", lambda *_: set())
+
+        adapter = APIServerAdapter(PlatformConfig(enabled=True))
+        monkeypatch.setattr(adapter, "_ensure_session_db", lambda: None)
+
+        agent = adapter._create_agent(session_id="api-session")
+
+        assert isinstance(agent, FakeAgent)
+        assert captured["model"] == "primary/model"
+
 
 # ---------------------------------------------------------------------------
 # Auth checking