Update app.py

app.py

@@ -43,12 +43,13 @@ client = InferenceClient(HF_MODEL_ID, token=os.getenv("HF_TOKEN"))
 #     return _llm
 
 llm = HuggingFaceEndpoint(
-    repo_id="ibm-granite/granite-3.1-8b-instruct", # Using Granite for 2026 stability
+    repo_id="meta-llama/Llama-3.1-8B-Instruct", # This is the "Live" 2026 model
+    # repo_id="ibm-granite/granite-3.1-8b-instruct", # Using Granite for 2026 stability
     task="text-generation",
     max_new_tokens=512,
     temperature=0.1,
     huggingfacehub_api_token=os.getenv("HF_TOKEN"),
-    wait_for_model=True
+    wait_for_model=True 
 )