Switch to llama-3.1-8b-instant (mixtral was deprecated)

app.py

@@ -360,8 +360,8 @@ def search_similar_questions(question: str) -> list:
 # Step 3  - Custom LLM
 from openai import OpenAI
 
-# Get model name from environment or use Groq's mixtral model
-LLM_MODEL = os.environ.get("LLM_MODEL", "mixtral-8x7b-32768")
+# Get model name from environment or use Groq's llama model (mixtral was deprecated)
+LLM_MODEL = os.environ.get("LLM_MODEL", "llama-3.1-8b-instant")
 
 def generate_stream(prompt, model=None):
     # Use environment variables for flexibility (OpenAI, Groq, or Custom HF Endpoint)