test: forward response mode in transport helper

tests/unit/test_openai_compat.py

@@ -101,10 +101,19 @@ async def _complete(
         reasoning_effort=reasoning_effort,
         first_token_event=first_token_event,
         content_event=content_event,
-        response_mode="text",
+        response_mode=response_mode,
     )
 
 
+async def test_complete_helper_forwards_response_mode():
+    """测试封装不得将显式保真模式静默退化成文本模式。"""
+    from unittest.mock import AsyncMock
+
+    transport = AsyncMock()
+    await _complete(transport, _source(), response_mode="openai")
+    assert transport.complete.call_args.kwargs["response_mode"] == "openai"
+
+
 # 单价刻意取"输出贵于输入"的真实形态: est_tokens 兜底把整估值塞进 completion
 # 时,虚高才显形(设计 §1 的 26 倍算例即此单价)。
 _PRICING = PricingTable({"qwen-max": ModelPrice(input_per_1m=1.0, output_per_1m=8.0)})