Try earlier llama model

app.py

@@ -53,7 +53,7 @@ def handle_submit():
             #model="Qwen/Qwen2.5-72B-Instruct",
             #model="Qwen/Qwen2.5-Coder-32B-Instruct",
             #model="mistralai/Mistral-7B-Instruct-v0.3",
-            model="meta-llama/Llama-3.2-3B",
+            model="meta-llama/Llama-3.1-8B-Instruct",
             messages=messages,
             max_tokens=1000
         )