diff --git a/profiler/src/profiler/TracyView.cpp b/profiler/src/profiler/TracyView.cpp
index de96d5d9..6865cbec 100644
--- a/profiler/src/profiler/TracyView.cpp
+++ b/profiler/src/profiler/TracyView.cpp
@@ -715,10 +715,17 @@ bool View::DrawImpl()
     auto& threadHints = m_worker.GetPendingThreadHints();
     if( !threadHints.empty() )
     {
+        m_threadReinsert.reserve( threadHints.size()  );
         for( auto v : threadHints )
         {
             auto it = std::find_if( m_threadOrder.begin(), m_threadOrder.end(), [v]( const auto& t ) { return t->id == v; } );
-            if( it != m_threadOrder.end() ) m_threadOrder.erase( it );      // Will be added in the correct place later, like any newly appearing thread
+            if( it != m_threadOrder.end() )
+            {
+                // Will be reinserted in the correct place later.
+                // A separate list is kept of threads that were already known to avoid having to figure out which one is missing in m_threadOrder.
+                m_threadReinsert.push_back( *it );
+                m_threadOrder.erase( it );
+            }
         }
         m_worker.ClearPendingThreadHints();
     }
diff --git a/profiler/src/profiler/TracyView.hpp b/profiler/src/profiler/TracyView.hpp
index 4855f1f8..6f7b76de 100644
--- a/profiler/src/profiler/TracyView.hpp
+++ b/profiler/src/profiler/TracyView.hpp
@@ -386,6 +386,7 @@ private:
     unordered_flat_map<const void*, int> m_gpuDrift;
     unordered_flat_map<const PlotData*, PlotView> m_plotView;
     Vector<const ThreadData*> m_threadOrder;
+    Vector<const ThreadData*> m_threadReinsert;
     Vector<float> m_threadDnd;
 
     tracy_force_inline bool& VisibleMsgThread( uint64_t thread )
diff --git a/profiler/src/profiler/TracyView_Timeline.cpp b/profiler/src/profiler/TracyView_Timeline.cpp
index 010ce1ed..107fce4b 100644
--- a/profiler/src/profiler/TracyView_Timeline.cpp
+++ b/profiler/src/profiler/TracyView_Timeline.cpp
@@ -366,11 +366,19 @@ void View::DrawTimeline()
         if( threadData.size() != m_threadOrder.size() )
         {
             m_threadOrder.reserve( threadData.size() );
-            for( size_t i=m_threadOrder.size(); i<threadData.size(); i++ )
+            // Only new threads are in the end of the worker's ThreadData vector.
+            // Threads which get reordered by received thread hints are not new, yet removed from m_threadOrder.
+            // Therefore, those are kept in the m_threadReinsert vector. As such, we will gather first threads from the
+            // reinsert vector, and afterwards the remaining ones must be new (and thus found at the end of threadData).
+            size_t numReinsert = m_threadReinsert.size();
+            size_t numNew = threadData.size() - m_threadOrder.size() - numReinsert;
+            for( size_t i = 0; i < numReinsert + numNew; i++ )
             {
-                auto it = std::upper_bound( m_threadOrder.begin(), m_threadOrder.end(), threadData[i]->groupHint, []( const auto& lhs, const auto& rhs ) { return lhs < rhs->groupHint; } );
-                m_threadOrder.insert( it, threadData[i] );
+                const ThreadData *td = i < numReinsert ? m_threadReinsert[i] : threadData[m_threadOrder.size()];
+                auto it = std::find_if( m_threadOrder.begin(), m_threadOrder.end(), [td]( const auto t ) { return td->groupHint < t->groupHint; } );
+                m_threadOrder.insert( it, td );
             }
+            m_threadReinsert.clear();
         }
         for( const auto& v : m_threadOrder )
         {