diff --git a/json/llm.json b/json/llm.json
index 0a4e63e..538c227 100644
--- a/json/llm.json
+++ b/json/llm.json
@@ -3,6 +3,7 @@
     "title": "大语言模型管理",
     "params": {
         "sortby": "created_at",
+        "logined_userorgid": "org_id",
         "browserfields": {
             "exclouded": ["id", "api_key"],
             "cwidth": {}
diff --git a/models/llm.json b/models/llm.json
index c4aee72..653ad5d 100644
--- a/models/llm.json
+++ b/models/llm.json
@@ -15,6 +15,7 @@
         {"name": "api_key", "title": "API密钥", "type": "str", "length": 500},
         {"name": "max_tokens", "title": "最大Token", "type": "int", "default": "8192"},
         {"name": "status", "title": "状态", "type": "str", "length": 20, "default": "active"},
+        {"name": "org_id", "title": "所属机构ID", "type": "str", "length": 32, "default": "0"},
         {"name": "description", "title": "描述", "type": "text"},
         {"name": "created_at", "title": "创建时间", "type": "timestamp"},
         {"name": "updated_at", "title": "更新时间", "type": "timestamp"}
diff --git a/models/pipelines.json b/models/pipelines.json
index 893f0a7..572dd8e 100644
--- a/models/pipelines.json
+++ b/models/pipelines.json
@@ -76,6 +76,13 @@
       "length": 500,
       "uitype": "text"
     },
+    {
+      "name": "default_model",
+      "title": "缺省模型",
+      "type": "str",
+      "length": 100,
+      "uitype": "code"
+    },
     {
       "name": "org_id",
       "title": "所属机构ID",
@@ -134,6 +141,13 @@
       "valuefield": "k",
       "textfield": "v",
       "cond": "parentid='pipeline_status'"
+    },
+    {
+      "field": "default_model",
+      "table": "llm",
+      "valuefield": "name",
+      "textfield": "name",
+      "cond": "status='active'"
     }
   ]
 }
\ No newline at end of file
diff --git a/pipeline_core/agent_config.py b/pipeline_core/agent_config.py
index 72776f4..432804d 100644
--- a/pipeline_core/agent_config.py
+++ b/pipeline_core/agent_config.py
@@ -360,21 +360,30 @@ async def load_agent_config(pipeline_id: str = None, project_id: str = None) ->
                         except (json.JSONDecodeError, TypeError):
                             pass
 
-            # 产线级配置
+            # 产线级配置 + 产线缺省模型
             if pipeline_id:
                 recs = await sor.sqlExe(
-                    "SELECT agent_config FROM pipelines WHERE id=${pid}$",
+                    "SELECT agent_config, default_model FROM pipelines WHERE id=${pid}$",
                     {"pid": pipeline_id},
                 )
                 if recs:
-                    raw = getattr(recs[0], "agent_config", "")
+                    raw = getattr(recs[0], "agent_config", "") or ""
+                    default_model = getattr(recs[0], "default_model", "") or ""
                     if raw:
                         try:
                             data = json.loads(raw) if isinstance(raw, str) else raw
                             data["tools"] = _merge_tools(data.get("tools", []), SDLC_DEFAULT_TOOLS)
+                            # 产线缺省模型：agent_config 未显式设 model_name 时用 default_model 兜底
+                            if not data.get("model_name") and default_model:
+                                data["model_name"] = default_model
                             return AgentConfig.from_dict(data)
                         except (json.JSONDecodeError, TypeError):
                             pass
+                    elif default_model:
+                        # agent_config 为空，仅产线设了缺省模型
+                        cfg = AgentConfig.from_dict(SDLC_DEFAULT_CONFIG.to_dict())
+                        cfg.model_name = default_model
+                        return cfg
     except Exception:
         pass