HARDFIX: Force llama-3.1-8b-instant model everywhere - 70B deprecated

src/api/agents/skill/orchestrator.py

@@ -1,6 +1,6 @@
 """Skill Agent Orchestrator"""
 from typing import Dict, Any, List, Optional
-from src.services.groq_service import generate_response
+from src.services.groq_service import generate_response, DEFAULT_MODEL
 from src.api.core.logger import logger
 
 
@@ -32,7 +32,8 @@ class SkillOrchestrator:
         ]
         
         try:
-            result = await generate_response(messages)
+            # Use DEFAULT_MODEL directly
+            result = await generate_response(messages, model=DEFAULT_MODEL)
             return {"skill_type": skill_type, "result": result, "status": "success"}
         except Exception as e:
             logger.error(f"Skill execution failed: {e}")

src/services/groq_service.py

@@ -8,12 +8,15 @@ from src.api.core.logger import logger
 
 client: Optional[Groq] = None
 
+# Default model - MUST use 8B as 70B is deprecated
+DEFAULT_MODEL = "llama-3.1-8b-instant"
+
 
 def get_groq_client() -> Groq:
     """Get or create Groq client instance"""
     global client
     if client is None:
-        client = Groq(api_key=settings.GROQ_API_KEY)
+        client = Groq(api_key=settings.GROQ_API_KEY or "dummy")
     return client
 
 
@@ -25,22 +28,16 @@ async def generate_response(
 ) -> str:
     """
     Generate response using Groq API
-    
-    Args:
-        messages: List of message dicts with 'role' and 'content'
-        model: Model name (defaults to settings.GROQ_MODEL)
-        temperature: Sampling temperature
-        max_tokens: Maximum tokens to generate
-    
-    Returns:
-        Generated response string
     """
     try:
         groq_client = get_groq_client()
-        model = model or settings.GROQ_MODEL
+        # Use DEFAULT_MODEL - 70b is deprecated!
+        model_name = model or getattr(settings, 'GROQ_MODEL', DEFAULT_MODEL) or DEFAULT_MODEL
+        
+        print(f"[DEBUG] Using model: {model_name}")
         
         response = groq_client.chat.completions.create(
-            model=model,
+            model=model_name,
             messages=messages,
             temperature=temperature,
             max_tokens=max_tokens,
@@ -61,10 +58,10 @@ async def stream_response(
     """Stream response from Groq API"""
     try:
         groq_client = get_groq_client()
-        model = model or settings.GROQ_MODEL
+        model_name = model or getattr(settings, 'GROQ_MODEL', DEFAULT_MODEL) or DEFAULT_MODEL
         
         stream = groq_client.chat.completions.create(
-            model=model,
+            model=model_name,
             messages=messages,
             temperature=temperature,
             stream=True,