diff --git a/app/api/v1/video_api.py b/app/api/v1/video_api.py
index 7cb6ba3..73a89bf 100644
--- a/app/api/v1/video_api.py
+++ b/app/api/v1/video_api.py
@@ -2,6 +2,7 @@
 视频操作专用 API 路由
 """
 
+import asyncio
 from typing import Optional
 from fastapi import APIRouter, Depends, Request
 from pydantic import BaseModel, Field
@@ -9,6 +10,7 @@ from pydantic import BaseModel, Field
 from app.services.grok.services.video import VideoService
 from app.core.auth import verify_api_key
 from app.core.logger import logger
+from app.core.exceptions import ValidationException
 from app.api.v1.chat import _build_streaming_response, _chat_error_as_success_response, _video_error_message
 
 router = APIRouter(tags=["Video"])
@@ -22,6 +24,8 @@ class VideoExtendRequest(BaseModel):
     aspect_ratio: Optional[str] = Field("16:9", description="视频比例")
     resolution: Optional[str] = Field("480p", description="分辨率 (480p/720p)")
     stream: Optional[bool] = Field(True, description="是否流式返回")
+    n: Optional[int] = Field(None, ge=1, le=4, description="生成数量 (1-4，仅非流式)")
+    concurrent: Optional[int] = Field(None, ge=1, le=4, description="并发视频数 (1-4，仅非流式)")
     video_extension_start_time: Optional[float] = Field(None, description="延长开始时间点")
     stitch_with_extend: Optional[bool] = Field(True, description="是否拼接之前的视频")
 
@@ -56,19 +60,67 @@ async def extend_video(request: VideoExtendRequest, raw_request: Request):
         # 即使留空，VideoService 也会处理，这里确保至少有一个 user 消息体以便提取
         messages.append({"role": "user", "content": ""})
 
-    try:
-        # 使用 VideoService.completions 统一入口
-        result = await VideoService.completions(
-            model=request.model,
-            messages=messages,
-            stream=request.stream,
-            aspect_ratio=request.aspect_ratio or "16:9",
-            video_length=video_length,
-            resolution=request.resolution or "480p",
-            extend_post_id=request.post_id,
-            video_extension_start_time=extension_start,
-            stitch_with_extend=request.stitch_with_extend
+    video_concurrent = int(request.n or request.concurrent or 1)
+    if request.stream and video_concurrent > 1:
+        raise ValidationException(
+            message="Streaming is only supported when n=1",
+            param="n",
+            code="invalid_stream_concurrent",
         )
+
+    try:
+        async def _run_single(single_stream: bool):
+            return await VideoService.completions(
+                model=request.model,
+                messages=messages,
+                stream=single_stream,
+                aspect_ratio=request.aspect_ratio or "16:9",
+                video_length=video_length,
+                resolution=request.resolution or "480p",
+                extend_post_id=request.post_id,
+                video_extension_start_time=extension_start,
+                stitch_with_extend=request.stitch_with_extend,
+            )
+
+        if video_concurrent <= 1:
+            result = await _run_single(bool(request.stream))
+        else:
+            results = await asyncio.gather(*[_run_single(False) for _ in range(video_concurrent)])
+            merged = dict(results[0]) if results else {
+                "id": "",
+                "object": "chat.completion",
+                "created": 0,
+                "model": request.model,
+                "choices": [],
+                "usage": {"prompt_tokens": 0, "completion_tokens": 0, "total_tokens": 0},
+            }
+
+            choices = []
+            usage_prompt = 0
+            usage_completion = 0
+            usage_total = 0
+            for idx, item in enumerate(results):
+                raw_choice = (
+                    item.get("choices", [{}])[0]
+                    if isinstance(item.get("choices"), list) and item.get("choices")
+                    else {}
+                )
+                choice = dict(raw_choice)
+                choice["index"] = idx
+                choices.append(choice)
+
+                usage = item.get("usage", {}) if isinstance(item, dict) else {}
+                usage_prompt += int(usage.get("prompt_tokens", 0) or 0)
+                usage_completion += int(usage.get("completion_tokens", 0) or 0)
+                usage_total += int(usage.get("total_tokens", 0) or 0)
+
+            merged["choices"] = choices
+            merged["usage"] = {
+                "prompt_tokens": usage_prompt,
+                "completion_tokens": usage_completion,
+                "total_tokens": usage_total,
+            }
+            result = merged
         
         if request.stream:
             return _build_streaming_response(result, raw_request, request.model)