diff --git a/apps/base/customer1/openclaw/configmap.yaml b/apps/base/customer1/openclaw/configmap.yaml
index 1b973ce..214710d 100644
--- a/apps/base/customer1/openclaw/configmap.yaml
+++ b/apps/base/customer1/openclaw/configmap.yaml
@@ -49,7 +49,7 @@ data:
       "agents": {
         "defaults": {
           "model": { 
-            "primary": "openai/Qwen/Qwen2.5-32B-Instruct",
+            "primary": "openai/solidrust/dolphin-2.9.2-qwen2-72b-AWQ",
             "fallbacks": ["google/gemini-3.1-pro-preview"]
           },
           "workspace": "~/.openclaw/workspace"
diff --git a/apps/base/customer1/openclaw/vllm-gemma.yaml b/apps/base/customer1/openclaw/vllm-gemma.yaml
index 9d94141..5f53b80 100644
--- a/apps/base/customer1/openclaw/vllm-gemma.yaml
+++ b/apps/base/customer1/openclaw/vllm-gemma.yaml
@@ -28,7 +28,9 @@ spec:
           command: ["python3", "-m", "vllm.entrypoints.openai.api_server"]
           args:
             - "--model"
-            - "Qwen/Qwen2.5-32B-Instruct" # Official Qwen 2.5 32B Base Model
+            - "solidrust/dolphin-2.9.2-qwen2-72b-AWQ" # Elite Uncensored 72B (AWQ Quantized for 80GB VRAM)
+            - "--quantization"
+            - "awq"
             - "--dtype"
             - "half"
             - "--max-model-len"