display model name

app.py

@@ -8,7 +8,12 @@ import requests
 import inspect
 import pandas as pd
 
-from agent import ANTHROPIC_MODEL_CHOICES, DEFAULT_ANTHROPIC_MODEL_ID, BasicAgent
+from agent import (
+    ANTHROPIC_MODEL_CHOICES,
+    DEFAULT_ANTHROPIC_MODEL_ID,
+    DEFAULT_MODEL_ID,
+    BasicAgent,
+)
 
 # (Keep Constants as is)
 # --- Constants ---
@@ -19,6 +24,15 @@ TOTAL_REQUESTS_PER_MINUTE = int(os.getenv("AGENT_REQUESTS_PER_MINUTE", "30"))
 _thread_local = threading.local()
 
 
+def describe_backend() -> str:
+    if os.getenv("ANTHROPIC_API_KEY"):
+        return f"**Backend:** Anthropic (`ANTHROPIC_API_KEY` found) — default model `{DEFAULT_ANTHROPIC_MODEL_ID}`"
+    return (
+        "**Backend:** Hugging Face Inference (no `ANTHROPIC_API_KEY` set) — default model "
+        f"`{os.getenv('HF_MODEL_ID') or DEFAULT_MODEL_ID}`"
+    )
+
+
 def get_worker_agent(model_id: str | None, requests_per_minute: int) -> BasicAgent:
     """One agent per worker thread: a CodeAgent's memory cannot be shared across runs in flight."""
     agent = getattr(_thread_local, "agent", None)
@@ -72,7 +86,7 @@ def run_and_submit_all( profile: gr.OAuthProfile | None, model_id: str, concurre
     worker_rpm = max(1, TOTAL_REQUESTS_PER_MINUTE // workers)
     try:
         # Fail fast on bad model configuration; workers build their own agents.
-        BasicAgent(model_id=model_id)
+        resolved_model = BasicAgent(model_id=model_id).model_id
     except Exception as e:
         print(f"Error instantiating agent: {e}")
         yield f"Error initializing agent: {e}", None
@@ -83,7 +97,7 @@ def run_and_submit_all( profile: gr.OAuthProfile | None, model_id: str, concurre
 
     # 2. Fetch Questions
     print(f"Fetching questions from: {questions_url}")
-    yield "Fetching questions...", None
+    yield f"Model: {resolved_model}\nFetching questions...", None
     try:
         response = requests.get(questions_url, timeout=15)
         response.raise_for_status()
@@ -119,8 +133,9 @@ def run_and_submit_all( profile: gr.OAuthProfile | None, model_id: str, concurre
     total = len(questions)
     rows: list[dict | None] = [None] * total
     started_all = time.monotonic()
-    print(f"Running agent on {total} questions with {workers} worker(s)...")
-    yield f"Running agent on {total} questions with {workers} worker(s)...", None
+    run_header = f"Model: {resolved_model} | {workers} parallel worker(s)"
+    print(f"Running agent on {total} questions. {run_header}")
+    yield f"{run_header}\nRunning agent on {total} questions...", None
 
     with ThreadPoolExecutor(max_workers=workers) as pool:
         futures = {
@@ -139,6 +154,7 @@ def run_and_submit_all( profile: gr.OAuthProfile | None, model_id: str, concurre
                 "Submitted Answer": display,
             }
             yield (
+                f"{run_header}\n"
                 f"Answered {done}/{total} questions "
                 f"({time.monotonic() - started_all:.0f}s elapsed)...",
                 pd.DataFrame([row for row in rows if row]),
@@ -167,6 +183,7 @@ def run_and_submit_all( profile: gr.OAuthProfile | None, model_id: str, concurre
         final_status = (
             f"Submission Successful!\n"
             f"User: {result_data.get('username')}\n"
+            f"Model: {resolved_model}\n"
             f"Overall Score: {result_data.get('score', 'N/A')}% "
             f"({result_data.get('correct_count', '?')}/{result_data.get('total_attempted', '?')} correct)\n"
             f"Message: {result_data.get('message', 'No message received.')}"
@@ -220,6 +237,8 @@ with gr.Blocks() as demo:
         """
     )
 
+    gr.Markdown(describe_backend())
+
     gr.LoginButton()
 
     model_dropdown = gr.Dropdown(