diff --git a/python/sglang/multimodal_gen/configs/sample/qwenimage.py b/python/sglang/multimodal_gen/configs/sample/qwenimage.py
index c24c1dd25..127dfa5fc 100644
--- a/python/sglang/multimodal_gen/configs/sample/qwenimage.py
+++ b/python/sglang/multimodal_gen/configs/sample/qwenimage.py
@@ -15,6 +15,13 @@ class QwenImageSamplingParams(SamplingParams):
     num_inference_steps: int = 50
 
 
+@dataclass
+class QwenImage2512SamplingParams(QwenImageSamplingParams):
+    negative_prompt: str = (
+        "低分辨率，低画质，肢体畸形，手指畸形，画面过饱和，蜡像感，人脸无细节，过度光滑，画面具有AI感。构图混乱。文字模糊，扭曲。"
+    )
+
+
 @dataclass
 class QwenImageEditPlusSamplingParams(QwenImageSamplingParams):
     # Denoising stage
diff --git a/python/sglang/multimodal_gen/registry.py b/python/sglang/multimodal_gen/registry.py
index b9e91d044..0979304c3 100644
--- a/python/sglang/multimodal_gen/registry.py
+++ b/python/sglang/multimodal_gen/registry.py
@@ -47,6 +47,7 @@ from sglang.multimodal_gen.configs.sample.hunyuan import (
     HunyuanSamplingParams,
 )
 from sglang.multimodal_gen.configs.sample.qwenimage import (
+    QwenImage2512SamplingParams,
     QwenImageEditPlusSamplingParams,
     QwenImageLayeredSamplingParams,
     QwenImageSamplingParams,
@@ -429,7 +430,12 @@ def _register_configs():
     register_configs(
         sampling_param_cls=QwenImageSamplingParams,
         pipeline_config_cls=QwenImagePipelineConfig,
-        hf_model_paths=["Qwen/Qwen-Image", "Qwen/Qwen-Image-2512"],
+        hf_model_paths=["Qwen/Qwen-Image"],
+    )
+    register_configs(
+        sampling_param_cls=QwenImage2512SamplingParams,
+        pipeline_config_cls=QwenImagePipelineConfig,
+        hf_model_paths=["Qwen/Qwen-Image-2512"],
     )
     register_configs(
         sampling_param_cls=QwenImageSamplingParams,