update configs (model routing)

src/config/settings.yaml

@@ -7,7 +7,7 @@ qdrant:
   # url: "https://2c6d0136-b6ca-4400-bac5-1703f58abc43.europe-west3-0.gcp.cloud.qdrant.io"
   # Current cluster with BAAI-bge-m3 and colSmol-500M-v2 collections:
   url: "${QDRANT_URL}"  # Load from environment variable (should be the us-west-1 cluster)
-  collection_name: "BAAI-bge-m3"  # Text-based RAG collection (41175 points)
+  collection_name: "BAAI-bge-m3-full"  # Full text-based RAG collection
   prefer_grpc: true
   api_key: "${QDRANT_API_KEY}"  # Load from environment variable
 
@@ -57,6 +57,17 @@ reader:
   OPENAI:
     model: "gpt-4o-mini"
 
+  OPENAI_STRONG:
+    model: "gpt-4.1"
+    temperature: 0.2
+
+  # Optional: dedicated model for final response generation.
+  # Uncomment and set to gpt-5 or similar when ready for highest-quality answers.
+  # Falls back to OPENAI_STRONG if not set.
+  # OPENAI_RESPONSE:
+  #   model: "gpt-4.1"
+  #   temperature: 0.3
+
   OLLAMA:
     model: "mistral-small3.1:24b-instruct-2503-q8_0"
     base_url: "http://10.1.4.192:11434/"