diff --git a/06_gpu_and_ml/tgi_mixtral.py b/06_gpu_and_ml/tgi_mixtral.py
index fd0d42784..7d61767e3 100644
--- a/06_gpu_and_ml/tgi_mixtral.py
+++ b/06_gpu_and_ml/tgi_mixtral.py
@@ -25,10 +25,13 @@
 
 GPU_CONFIG = gpu.A100(memory=40, count=4)
 MODEL_ID = "mistralai/Mixtral-8x7B-Instruct-v0.1"
+MODEL_REVISION = "f1ca00645f0b1565c7f9a1c863d2be6ebf896b04"
 # Add `["--quantize", "gptq"]` for TheBloke GPTQ models.
 LAUNCH_FLAGS = [
     "--model-id",
     MODEL_ID,
+    "--revision",
+    MODEL_REVISION,
     "--port",
     "8000",
 ]
@@ -52,6 +55,8 @@ def download_model():
             "text-generation-server",
             "download-weights",
             MODEL_ID,
+            "--revision",
+            MODEL_REVISION,
         ]
     )