feat: carry the per-call tier down to the transport that must send it

The Transport port took the request apart into five arguments, so a tier
placed on ChatRequest could never reach _build_payload: the field was set,
read by nobody, and silently ignored - the exact shape of failure that sent
downstream to extra_body in the first place.

complete() now takes reasoning_effort with no default, matching the
TelemetryRecorder convention: a default would turn a missing hand-off into
a silent 'no opinion'. All four fakes move with it, since @runtime_checkable
checks method names and not signatures.

EmbeddingTransport and OcrTransport are deliberately left alone - they have
no reasoning semantics - and a test now holds that line.

_build_payload drops its inline sugar conversion for effective_effort(), so
the guard and the hot path share one judgement, and passes the source's
effort_fallback for the same reason.

tests/unit/test_client.py

@@ -24,6 +24,7 @@ from polygateway.transports.openai_compat import OpenAICompatTransport
 from polygateway.types import (
     BackpressurePolicy,
     BreakerConfig,
+    Effort,
     GlobalLimits,
     RetryPolicy,
     SourceConfig,
@@ -219,6 +220,31 @@ class TestReasoningEffortPriority:
         overrides.setdefault("model", "glm-5.3")
         return _source(provider="zhipu", **overrides)
 
+    async def test_request_effort_wins_over_source(self):
+        captured = []
+        source = self._zhipu(reasoning_effort=Effort.LOW)
+        async with self._capturing_client(captured, sources=[source]) as client:
+            await client.chat([{"role": "user", "content": "hi"}], reasoning_effort=Effort.MAX)
+        assert captured[0]["reasoning_effort"] == "max"
+        assert captured[0]["thinking"] == {"type": "enabled"}
+
+    async def test_none_request_does_not_clear_source(self):
+        """请求级不表态 ≠ 请求级要求"不推理": 前者必须让源级默认继续生效。"""
+        captured = []
+        source = self._zhipu(reasoning_effort=Effort.LOW)
+        async with self._capturing_client(captured, sources=[source]) as client:
+            await client.chat([{"role": "user", "content": "hi"}])
+        assert captured[0]["reasoning_effort"] == "low"
+
+    async def test_request_none_tier_is_an_opinion_not_an_absence(self):
+        """请求级 `none` 是"要求不推理",不得被当成"没表态"而回落到源级档位。"""
+        captured = []
+        source = self._zhipu(model="glm-5.2", reasoning_effort=Effort.MAX)
+        async with self._capturing_client(captured, sources=[source]) as client:
+            await client.chat([{"role": "user", "content": "hi"}], reasoning_effort=Effort.NONE)
+        assert captured[0]["thinking"] == {"type": "disabled"}
+        assert "reasoning_effort" not in captured[0]
+
     @pytest.mark.parametrize(
         ("provider", "model", "fragment"),
         [