diff --git a/gateway/slash_commands_model.py b/gateway/slash_commands_model.py
index 1491e62fb675d..f1013a9d58051 100644
--- a/gateway/slash_commands_model.py
+++ b/gateway/slash_commands_model.py
@@ -11,7 +11,9 @@
 import asyncio
 import contextlib
 import dataclasses
+import json
 import logging
+from types import SimpleNamespace
 from typing import Any, Optional
 
 from agent.i18n import t
@@ -34,6 +36,29 @@
 # /reasoning display-toggle arguments -> show_reasoning value.
 _REASONING_DISPLAY_TOGGLES = {"show": True, "on": True, "hide": False, "off": False}
 
+def _configured_provider_routing_line(routing: dict, model: str) -> Optional[str]:
+    """Render validated request routing without implying which provider served."""
+    from agent.chat_completion_helpers import _provider_preferences_for_agent
+
+    if not isinstance(routing, dict):
+        return None
+    effective = _provider_preferences_for_agent(SimpleNamespace(
+        model=model,
+        providers_allowed=routing.get("only"),
+        providers_ignored=routing.get("ignore"),
+        providers_order=routing.get("order"),
+        provider_sort=routing.get("sort"),
+        provider_require_parameters=routing.get("require_parameters"),
+        provider_data_collection=routing.get("data_collection"),
+    ))
+    if not effective:
+        return None
+    try:
+        encoded = json.dumps(effective, ensure_ascii=False, separators=(",", ":"))
+    except (TypeError, ValueError):
+        return None
+    return f"provider_routing: {encoded}"
+
 
 def _model_switch_skew_guard() -> Optional[str]:
     """Refuse a model switch when the gateway is running stale code: a first-time lazy import on
@@ -273,14 +298,26 @@ async def _model_switch_confirmation(
         ]
         # Provider-aware chain: Codex OAuth, Copilot and Nous caps win over the raw models.dev entry.
         mi = result.model_info
+        gateway_cfg: dict = {}
         model_cfg: dict = {}
         config_ctx = None
         with contextlib.suppress(Exception):  # fail-open on config read errors
-            model_cfg = _load_gateway_config().get("model", {})
+            gateway_cfg = _load_gateway_config()
+            model_cfg = gateway_cfg.get("model", {})
             if isinstance(model_cfg, dict) and model_cfg.get("context_length") is not None:
                 config_ctx = int(model_cfg["context_length"])
         if not isinstance(model_cfg, dict):
             model_cfg = {}
+        if (
+            result.target_provider == "openrouter"
+            or base_url_host_matches(result.base_url or "", "openrouter.ai")
+        ):
+            routing_line = _configured_provider_routing_line(
+                getattr(self, "_provider_routing", {}),
+                result.new_model,
+            )
+            if routing_line:
+                lines.append(routing_line)
         ctx_len = await resolve_display_context_length_async(
             result.new_model, result.target_provider,
             base_url=result.base_url or ctx.current_base_url or "",
diff --git a/tests/gateway/test_model_switch_provider_routing_display.py b/tests/gateway/test_model_switch_provider_routing_display.py
new file mode 100644
index 0000000000000..79cfdde636ab5
--- /dev/null
+++ b/tests/gateway/test_model_switch_provider_routing_display.py
@@ -0,0 +1,105 @@
+"""Gateway model-switch confirmations expose configured OpenRouter routing.
+
+Regression for #110784.
+"""
+
+from datetime import date
+from unittest.mock import AsyncMock, MagicMock
+
+import pytest
+
+from gateway.config import Platform
+from gateway.platforms.event import MessageEvent, MessageType
+from gateway.run import GatewayRunner
+from gateway.session import SessionSource
+from gateway.slash_commands_model import _configured_provider_routing_line
+
+
+@pytest.mark.parametrize("routing", [{"sort": "bogus"}, {"only": [date(2026, 1, 1)]}])
+def test_routing_display_omits_values_the_request_path_cannot_emit(monkeypatch, routing):
+    monkeypatch.setattr("hermes_cli.config.load_config_readonly", lambda: {})
+
+    assert _configured_provider_routing_line(routing, "test/model") is None
+
+
+@pytest.mark.asyncio
+async def test_model_command_reads_routing_from_profile_config(tmp_path, monkeypatch):
+    import gateway.run as gateway_run
+    from hermes_cli.model_switch import ModelSwitchResult
+
+    hermes_home = tmp_path / ".hermes"
+    hermes_home.mkdir()
+    (hermes_home / "config.yaml").write_text(
+        """
+model:
+  default: old-model
+  provider: openrouter
+provider_routing:
+  sort: price
+  models:
+    "z-ai/glm-5.3:exacto":
+      only: [fireworks]
+      data_collection: deny
+""".lstrip(),
+        encoding="utf-8",
+    )
+    monkeypatch.setattr(gateway_run, "_hermes_home", hermes_home)
+    monkeypatch.setattr("hermes_constants.get_hermes_home", lambda: hermes_home)
+    monkeypatch.setattr("hermes_cli.config.get_hermes_home", lambda: hermes_home)
+    monkeypatch.setattr("agent.models_dev.fetch_models_dev", lambda: {})
+    monkeypatch.setattr(
+        "hermes_cli.model_cost_guard.expensive_model_warning",
+        lambda *_args, **_kwargs: None,
+    )
+    monkeypatch.setattr(
+        "hermes_cli.model_switch.switch_model",
+        lambda **_kwargs: ModelSwitchResult(
+            success=True,
+            new_model="z-ai/glm-5.3:exacto",
+            target_provider="openrouter",
+            provider_changed=False,
+            api_key="test",
+            base_url="https://openrouter.ai/api/v1",
+            api_mode="chat_completions",
+            provider_label="OpenRouter",
+        ),
+    )
+
+    async def _no_context(*_args, **_kwargs):
+        return None
+
+    monkeypatch.setattr(
+        "hermes_cli.model_switch.resolve_display_context_length_async",
+        _no_context,
+    )
+
+    runner = object.__new__(GatewayRunner)
+    runner.adapters = {}
+    runner._voice_mode = {}
+    runner._session_model_overrides = {}
+    runner._pending_one_turn_model_restores = {}
+    runner._running_agents = {}
+    runner._provider_routing = {"sort": " Throughput "}
+    store = MagicMock()
+    store.set_model_override = AsyncMock()
+    store._store = None
+    setattr(runner, "session_store", None)
+    runner._async_session_store = store
+    event = MessageEvent(
+        text="/model glm53",
+        message_type=MessageType.TEXT,
+        source=SessionSource(
+            platform=Platform.SLACK,
+            user_id="user",
+            chat_id="channel",
+            chat_type="group",
+        ),
+    )
+
+    confirmation = await runner._handle_model_command(event)
+
+    assert confirmation is not None
+    assert (
+        'provider_routing: {"only":["fireworks"],"sort":"throughput",'
+        '"data_collection":"deny"}' in confirmation
+    )