diff --git a/llama_stack/distribution/routers/routers.py b/llama_stack/distribution/routers/routers.py
index c637a5b23..1a95ad45b 100644
--- a/llama_stack/distribution/routers/routers.py
+++ b/llama_stack/distribution/routers/routers.py
@@ -190,23 +190,18 @@ class InferenceRouter(Inference):
             )
         return metric_events
 
-    async def _add_token_metrics(
+    async def _compute_and_log_token_usage(
         self,
         prompt_tokens: int,
         completion_tokens: int,
         total_tokens: int,
         model: Model,
-        target: Any,
-    ) -> None:
-        metrics = getattr(target, "metrics", None)
-        if metrics is None:
-            target.metrics = []
-
+    ) -> List[MetricEvent]:
         metrics = self._construct_metrics(prompt_tokens, completion_tokens, total_tokens, model)
-        target.metrics.extend(metrics)
         if self.telemetry:
             for metric in metrics:
                 await self.telemetry.log_event(metric)
+        return metrics
 
     async def _count_tokens(
         self,
@@ -296,13 +291,13 @@ class InferenceRouter(Inference):
                             tool_config.tool_prompt_format,
                         )
                         total_tokens = (prompt_tokens or 0) + (completion_tokens or 0)
-                        await self._add_token_metrics(
+                        metrics = await self._compute_and_log_token_usage(
                             prompt_tokens or 0,
                             completion_tokens or 0,
                             total_tokens,
                             model,
-                            chunk,
                         )
+                        chunk.metrics = metrics if chunk.metrics is None else chunk.metrics + metrics
                     yield chunk
 
             return stream_generator()
@@ -313,13 +308,13 @@ class InferenceRouter(Inference):
                 tool_config.tool_prompt_format,
             )
             total_tokens = (prompt_tokens or 0) + (completion_tokens or 0)
-            await self._add_token_metrics(
+            metrics = await self._compute_and_log_token_usage(
                 prompt_tokens or 0,
                 completion_tokens or 0,
                 total_tokens,
                 model,
-                response,
             )
+            response.metrics = metrics if response.metrics is None else response.metrics + metrics
             return response
 
     async def completion(
@@ -362,13 +357,13 @@ class InferenceRouter(Inference):
                     if hasattr(chunk, "stop_reason") and chunk.stop_reason and self.telemetry:
                         completion_tokens = await self._count_tokens(completion_text)
                         total_tokens = (prompt_tokens or 0) + (completion_tokens or 0)
-                        await self._add_token_metrics(
+                        metrics = await self._compute_and_log_token_usage(
                             prompt_tokens or 0,
                             completion_tokens or 0,
                             total_tokens,
                             model,
-                            chunk,
                         )
+                        chunk.metrics = metrics if chunk.metrics is None else chunk.metrics + metrics
                     yield chunk
 
             return stream_generator()
@@ -376,13 +371,13 @@ class InferenceRouter(Inference):
             response = await provider.completion(**params)
             completion_tokens = await self._count_tokens(response.content)
             total_tokens = (prompt_tokens or 0) + (completion_tokens or 0)
-            await self._add_token_metrics(
+            metrics = await self._compute_and_log_token_usage(
                 prompt_tokens or 0,
                 completion_tokens or 0,
                 total_tokens,
                 model,
-                response,
             )
+            response.metrics = metrics if response.metrics is None else response.metrics + metrics
             return response
 
     async def embeddings(