diff --git a/crates/common/src/tokenizer.rs b/crates/common/src/tokenizer.rs
index 11ce7295..198c2af7 100644
--- a/crates/common/src/tokenizer.rs
+++ b/crates/common/src/tokenizer.rs
@@ -12,10 +12,15 @@ pub fn token_count(model_name: &str, text: &str) -> Result<usize, String> {
                 "tiktoken_rs: unsupported model: {}, using gpt-4 to compute token count",
                 model_name
             );
-
             "gpt-4"
         }
-        true => model_name,
+        true => {
+            if model_name.starts_with("gpt-4.1") {
+                "gpt-4o"
+            } else {
+                model_name
+            }
+        }
     };
 
     // Consideration: is it more expensive to instantiate the BPE object every time, or to contend the singleton?
diff --git a/demos/use_cases/preference_based_routing/arch_config.yaml b/demos/use_cases/preference_based_routing/arch_config.yaml
index 682527ca..c01f3ffc 100644
--- a/demos/use_cases/preference_based_routing/arch_config.yaml
+++ b/demos/use_cases/preference_based_routing/arch_config.yaml
@@ -1,7 +1,7 @@
 version: "0.1-beta"
 
 routing:
-  model: gpt-4o
+  model: archgw-v1-router-model
 
 listeners:
   egress_traffic:
@@ -14,7 +14,7 @@ llm_providers:
 
   - name: archgw-v1-router-model
     provider_interface: openai
-    model: cotran2/llama-4-epoch
+    model: cotran2/qwen-4-epoch-2600
     base_url: http://34.46.85.85:8000/v1
 
   - name: gpt-4o
diff --git a/demos/use_cases/preference_based_routing/docker-compose.yaml b/demos/use_cases/preference_based_routing/docker-compose.yaml
index c2d794c6..54158f73 100644
--- a/demos/use_cases/preference_based_routing/docker-compose.yaml
+++ b/demos/use_cases/preference_based_routing/docker-compose.yaml
@@ -1,17 +1,14 @@
 services:
 
-  chatbot_ui:
-    build:
-      context: ../../shared/chatbot_ui
-      dockerfile: Dockerfile
+  open-web-ui:
+    image: ghcr.io/open-webui/open-webui:main
+    restart: always
     ports:
-      - "18080:8080"
+      - "8080:8080"
     environment:
-      - CHAT_COMPLETION_ENDPOINT=http://host.docker.internal:12000/v1
-    extra_hosts:
-      - "host.docker.internal:host-gateway"
-    volumes:
-      - ./arch_config.yaml:/app/arch_config.yaml
+      - DEFAULT_MODEL=gpt-4o-mini
+      - ENABLE_OPENAI_API=true
+      - OPENAI_API_BASE_URL=http://host.docker.internal:12000/v1
 
   jaeger:
     build:
diff --git a/demos/use_cases/preference_based_routing/test_router_endpoint.rest b/demos/use_cases/preference_based_routing/test_router_endpoint.rest
index 9fc6f6fe..d4b947c8 100644
--- a/demos/use_cases/preference_based_routing/test_router_endpoint.rest
+++ b/demos/use_cases/preference_based_routing/test_router_endpoint.rest
@@ -1,10 +1,10 @@
 @arch_llm_router_endpoint = http://35.192.87.187:8000
 
-POST {{arch_llm_router_endpoint}}/v1/chat/completions HTTP/1.1
+POST http://34.46.85.85:8000/v1/chat/completions HTTP/1.1
 Content-Type: application/json
 
 {
-  "model": "cotran2/llama-1b-4-26",
+  "model": "cotran2/qwen-4-epoch-2600",
   "messages": [
     {
       "role": "user",