diff --git a/CHANGELOG.md b/CHANGELOG.md
index fd24e6b7fef..205aa7ea3d6 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -3,6 +3,7 @@
 * [FEATURE] TraceQL support for event scope and event:name intrinsic [#3708](https://github.com/grafana/tempo/pull/3708) (@stoewer)
 * [FEATURE] Flush and query RF1 blocks for TraceQL metric queries [#3628](https://github.com/grafana/tempo/pull/3628) [#3691](https://github.com/grafana/tempo/pull/3691) [#3723](https://github.com/grafana/tempo/pull/3723) (@mapno)
 * [ENHANCEMENT] Tag value lookup use protobuf internally for improved latency [#3731](https://github.com/grafana/tempo/pull/3731) (@mdisibio)
+* [ENHANCEMENT] TraceQL metrics queries use protobuf internally for improved latency [#3745](https://github.com/grafana/tempo/pull/3745) (@mdisibio)
 * [ENHANCEMENT] Improve use of OTEL semantic conventions on the service graph [#3711](https://github.com/grafana/tempo/pull/3711) (@zalegrala)
 * [ENHANCEMENT] Performance improvement for `rate() by ()` queries [#3719](https://github.com/grafana/tempo/pull/3719) (@mapno)
 * [BUGFIX] Fix metrics queries when grouping by attributes that may not exist [#3734](https://github.com/grafana/tempo/pull/3734) (@mdisibio)
diff --git a/modules/frontend/metrics_query_range_sharder.go b/modules/frontend/metrics_query_range_sharder.go
index 96a66963b9f..65eae768421 100644
--- a/modules/frontend/metrics_query_range_sharder.go
+++ b/modules/frontend/metrics_query_range_sharder.go
@@ -389,6 +389,7 @@ func (s *queryRangeSharder) buildBackendRequests(ctx context.Context, tenantID s
 			queryRangeReq.End += queryRangeReq.Step
 
 			subR = api.BuildQueryRangeRequest(subR, queryRangeReq)
+			subR.Header.Set(api.HeaderAccept, api.HeaderAcceptProtobuf)
 
 			prepareRequestForQueriers(subR, tenantID, subR.URL.Path, subR.URL.Query())
 			// TODO: Handle sampling rate
@@ -438,7 +439,10 @@ func (s *queryRangeSharder) generatorRequest(searchReq tempopb.QueryRangeRequest
 
 	searchReq.QueryMode = querier.QueryModeRecent
 
-	return s.toUpstreamRequest(parent.Context(), searchReq, parent, tenantID)
+	req := s.toUpstreamRequest(parent.Context(), searchReq, parent, tenantID)
+	req.Header.Set(api.HeaderAccept, api.HeaderAcceptProtobuf)
+
+	return req
 }
 
 func (s *queryRangeSharder) toUpstreamRequest(ctx context.Context, req tempopb.QueryRangeRequest, parent *http.Request, tenantID string) *http.Request {
diff --git a/modules/querier/http.go b/modules/querier/http.go
index 301aae491f1..fb67cdd78f1 100644
--- a/modules/querier/http.go
+++ b/modules/querier/http.go
@@ -409,22 +409,7 @@ func (q *Querier) QueryRangeHandler(w http.ResponseWriter, r *http.Request) {
 			return
 		}
 
-		m := jsonpb.Marshaler{}
-		jsBytes, funcErr := m.MarshalToString(resp)
-		if funcErr != nil {
-			errHandler(ctx, span, funcErr)
-			http.Error(w, funcErr.Error(), http.StatusInternalServerError)
-			return
-		}
-
-		_, funcErr = w.Write([]byte(jsBytes))
-		if funcErr != nil {
-			errHandler(ctx, span, funcErr)
-			http.Error(w, funcErr.Error(), http.StatusInternalServerError)
-			return
-		}
-
-		w.Header().Set(api.HeaderContentType, api.HeaderAcceptJSON)
+		writeFormattedContentForRequest(w, r, resp)
 	}()
 
 	req, err := api.ParseQueryRangeRequest(r)