diff --git a/src/swarm/manager.py b/src/swarm/manager.py
index 579efef..9e3f146 100644
--- a/src/swarm/manager.py
+++ b/src/swarm/manager.py
@@ -357,13 +357,21 @@ class SwarmManager:
         if not self.mcp_mode:
             print(f"🔄 Starting stream from {fastest_worker.name}...")
         chunk_count = 0
+        total_chars = 0
+        start_time = asyncio.get_event_loop().time()
         async for chunk in fastest_worker.generate_with_progress_stream(request):
             chunk_count += 1
+            total_chars += len(chunk)
             if not self.mcp_mode and chunk_count % 50 == 0:  # Print progress every 50 chunks
                 print(f"   Streamed {chunk_count} chunks...")
             yield chunk
+        end_time = asyncio.get_event_loop().time()
+        duration = end_time - start_time
+        # Estimate tokens (roughly 4 chars per token)
+        estimated_tokens = total_chars // 4
+        tps = estimated_tokens / duration if duration > 0 else 0
         if not self.mcp_mode:
-            print(f"   Stream complete: {chunk_count} chunks total")
+            print(f"   Stream complete: {chunk_count} chunks, {estimated_tokens} tokens, {tps:.1f} t/s")
     
     def get_status(self) -> SwarmStatus:
         """Get current swarm status."""
diff --git a/src/swarm/worker.py b/src/swarm/worker.py
index 09b877b..3eb9fa7 100644
--- a/src/swarm/worker.py
+++ b/src/swarm/worker.py
@@ -158,11 +158,6 @@ class SwarmWorker:
                 prompt_tokens = len(request.prompt) // 4
             
             self._context_used = min(prompt_tokens + request.max_tokens, 131072)  # Cap at reasonable max
-            
-            # Debug: Log the actual prompt size received
-            import logging
-            logger = logging.getLogger(__name__)
-            logger.info(f"🔍 WORKER {self.name}: Received prompt with {len(request.prompt)} chars, ~{prompt_tokens} tokens + {request.max_tokens} max = {self._context_used} context")
         
         try:
             start_time = time.time()