From a16ff2ea883b0cecd8c4ba3306b962c0ef3b8a68 Mon Sep 17 00:00:00 2001
From: Hyungi Ahn <hyungiahn@Hyungiui-MacBookPro.local>
Date: Mon, 6 Apr 2026 12:52:31 +0900
Subject: [PATCH] =?UTF-8?q?fix:=20max=5Ftokens=20=EC=B6=94=EA=B0=80=20?=
 =?UTF-8?q?=E2=80=94=20Gemma=2016000,=20EXAONE=204096?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

응답이 중간에 끊기는 문제 해결. ModelAdapter에 max_tokens
파라미터 추가, stream/complete 양쪽 payload에 반영.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
---
 nanoclaude/services/backend_registry.py | 1 +
 nanoclaude/services/model_adapter.py    | 4 ++++
 2 files changed, 5 insertions(+)

diff --git a/nanoclaude/services/backend_registry.py b/nanoclaude/services/backend_registry.py
index ac196b7..b72fcfb 100644
--- a/nanoclaude/services/backend_registry.py
+++ b/nanoclaude/services/backend_registry.py
@@ -58,6 +58,7 @@ class BackendRegistry:
             system_prompt=REASONER_PROMPT,
             temperature=settings.reasoning_temperature,
             timeout=settings.reasoning_timeout,
+            max_tokens=16000,
         )
 
     def start_health_loop(self, interval: float = 30.0) -> None:
diff --git a/nanoclaude/services/model_adapter.py b/nanoclaude/services/model_adapter.py
index 5b289c2..9690709 100644
--- a/nanoclaude/services/model_adapter.py
+++ b/nanoclaude/services/model_adapter.py
@@ -23,6 +23,7 @@ class ModelAdapter:
         system_prompt: str,
         temperature: float = 0.7,
         timeout: float = 120.0,
+        max_tokens: int = 4096,
     ):
         self.name = name
         self.base_url = base_url
@@ -30,6 +31,7 @@ class ModelAdapter:
         self.system_prompt = system_prompt
         self.temperature = temperature
         self.timeout = timeout
+        self.max_tokens = max_tokens
 
     async def stream_chat(self, message: str, *, messages: list[dict] | None = None) -> AsyncGenerator[str, None]:
         """스트리밍 호출. content chunk를 yield. messages 직접 전달 가능."""
@@ -43,6 +45,7 @@ class ModelAdapter:
             "messages": messages,
             "stream": True,
             "temperature": self.temperature,
+            "max_tokens": self.max_tokens,
         }
 
         async with httpx.AsyncClient(timeout=self.timeout) as client:
@@ -92,6 +95,7 @@ class ModelAdapter:
             "messages": messages,
             "stream": False,
             "temperature": self.temperature,
+            "max_tokens": self.max_tokens,
         }
 
         async with httpx.AsyncClient(timeout=self.timeout) as client: