diff --git a/gpt4all-chat/chat.cpp b/gpt4all-chat/chat.cpp
index a57f4997..f33a77fd 100644
--- a/gpt4all-chat/chat.cpp
+++ b/gpt4all-chat/chat.cpp
@@ -3,6 +3,7 @@
 #include "localdocs.h"
 #include "network.h"
 #include "download.h"
+#include "server.h"
 
 Chat::Chat(QObject *parent)
     : QObject(parent)
@@ -53,7 +54,7 @@ void Chat::connectLLM()
     connect(m_llmodel, &ChatLLM::promptProcessing, this, &Chat::promptProcessing, Qt::QueuedConnection);
     connect(m_llmodel, &ChatLLM::responseStopped, this, &Chat::responseStopped, Qt::QueuedConnection);
     connect(m_llmodel, &ChatLLM::modelNameChanged, this, &Chat::handleModelNameChanged, Qt::QueuedConnection);
-    connect(m_llmodel, &ChatLLM::modelLoadingError, this, &Chat::modelLoadingError, Qt::QueuedConnection);
+    connect(m_llmodel, &ChatLLM::modelLoadingError, this, &Chat::handleModelLoadingError, Qt::QueuedConnection);
     connect(m_llmodel, &ChatLLM::recalcChanged, this, &Chat::handleRecalculating, Qt::QueuedConnection);
     connect(m_llmodel, &ChatLLM::generatedNameChanged, this, &Chat::generatedNameChanged, Qt::QueuedConnection);
 
@@ -243,6 +244,8 @@ void Chat::setModelName(const QString &modelName)
 {
     // doesn't block but will unload old model and load new one which the gui can see through changes
     // to the isModelLoaded property
+    m_modelLoadingError = QString();
+    emit modelLoadingErrorChanged();
     emit modelNameChangeRequested(modelName);
 }
 
@@ -270,11 +273,15 @@ bool Chat::isRecalc() const
 
 void Chat::loadDefaultModel()
 {
+    m_modelLoadingError = QString();
+    emit modelLoadingErrorChanged();
     emit loadDefaultModelRequested();
 }
 
 void Chat::loadModel(const QString &modelName)
 {
+    m_modelLoadingError = QString();
+    emit modelLoadingErrorChanged();
     emit loadModelRequested(modelName);
 }
 
@@ -322,6 +329,13 @@ void Chat::handleModelNameChanged()
     emit modelNameChanged();
 }
 
+void Chat::handleModelLoadingError(const QString &error)
+{
+    qWarning() << "ERROR:" << qPrintable(error) << "id" << id();
+    m_modelLoadingError = error;
+    emit modelLoadingErrorChanged();
+}
+
 bool Chat::serialize(QDataStream &stream, int version) const
 {
     stream << m_creationDate;
diff --git a/gpt4all-chat/chat.h b/gpt4all-chat/chat.h
index 72dcccdb..7d6ea593 100644
--- a/gpt4all-chat/chat.h
+++ b/gpt4all-chat/chat.h
@@ -8,7 +8,6 @@
 #include "chatllm.h"
 #include "chatmodel.h"
 #include "database.h"
-#include "server.h"
 
 class Chat : public QObject
 {
@@ -25,6 +24,7 @@ class Chat : public QObject
     Q_PROPERTY(bool isServer READ isServer NOTIFY isServerChanged)
     Q_PROPERTY(QString responseState READ responseState NOTIFY responseStateChanged)
     Q_PROPERTY(QList<QString> collectionList READ collectionList NOTIFY collectionListChanged)
+    Q_PROPERTY(QString modelLoadingError READ modelLoadingError NOTIFY modelLoadingErrorChanged)
     QML_ELEMENT
     QML_UNCREATABLE("Only creatable from c++!")
 
@@ -89,6 +89,8 @@ public:
     Q_INVOKABLE void removeCollection(const QString &collection);
     void resetResponseState();
 
+    QString modelLoadingError() const { return m_modelLoadingError; }
+
 public Q_SLOTS:
     void serverNewPromptResponsePair(const QString &prompt);
 
@@ -113,7 +115,7 @@ Q_SIGNALS:
     void loadModelRequested(const QString &modelName);
     void generateNameRequested();
     void modelListChanged();
-    void modelLoadingError(const QString &error);
+    void modelLoadingErrorChanged();
     void isServerChanged();
     void collectionListChanged();
 
@@ -125,12 +127,14 @@ private Q_SLOTS:
     void generatedNameChanged();
     void handleRecalculating();
     void handleModelNameChanged();
+    void handleModelLoadingError(const QString &error);
 
 private:
     QString m_id;
     QString m_name;
     QString m_userName;
     QString m_savedModelName;
+    QString m_modelLoadingError;
     QList<QString> m_collections;
     ChatModel *m_chatModel;
     bool m_responseInProgress;
diff --git a/gpt4all-chat/chatlistmodel.cpp b/gpt4all-chat/chatlistmodel.cpp
index a794284d..24e9a9bb 100644
--- a/gpt4all-chat/chatlistmodel.cpp
+++ b/gpt4all-chat/chatlistmodel.cpp
@@ -233,7 +233,7 @@ void ChatListModel::restoreChat(Chat *chat)
 {
     chat->setParent(this);
     connect(chat, &Chat::nameChanged, this, &ChatListModel::nameChanged);
-    connect(chat, &Chat::modelLoadingError, this, &ChatListModel::handleModelLoadingError);
+    connect(chat, &Chat::modelLoadingErrorChanged, this, &ChatListModel::handleModelLoadingError);
 
     if (m_dummyChat) {
         beginResetModel();
diff --git a/gpt4all-chat/chatlistmodel.h b/gpt4all-chat/chatlistmodel.h
index 106c2fcb..bc38a414 100644
--- a/gpt4all-chat/chatlistmodel.h
+++ b/gpt4all-chat/chatlistmodel.h
@@ -227,10 +227,9 @@ private Q_SLOTS:
         emit dataChanged(index, index, {NameRole});
     }
 
-    void handleModelLoadingError(const QString &error)
+    void handleModelLoadingError()
     {
         Chat *chat = qobject_cast<Chat *>(sender());
-        qWarning() << "ERROR:" << qPrintable(error) << "id" << chat->id();
         removeChat(chat);
     }
 
diff --git a/gpt4all-chat/chatllm.cpp b/gpt4all-chat/chatllm.cpp
index 7666fddc..d4c8269e 100644
--- a/gpt4all-chat/chatllm.cpp
+++ b/gpt4all-chat/chatllm.cpp
@@ -228,6 +228,11 @@ bool ChatLLM::loadModel(const QString &modelName)
                 case 'M': m_modelType = LLModelType::MPT_; break;
                 default: delete std::exchange(m_modelInfo.model, nullptr);
                 }
+            } else {
+                if (!m_isServer)
+                    LLModelStore::globalInstance()->releaseModel(m_modelInfo); // release back into the store
+                m_modelInfo = LLModelInfo();
+                emit modelLoadingError(QString("Could not load model due to invalid format for %1").arg(modelName));
             }
         }
 #if defined(DEBUG_MODEL_LOADING)
@@ -249,8 +254,8 @@ bool ChatLLM::loadModel(const QString &modelName)
     } else {
         if (!m_isServer)
             LLModelStore::globalInstance()->releaseModel(m_modelInfo); // release back into the store
-        const QString error = QString("Could not find model %1").arg(modelName);
-        emit modelLoadingError(error);
+        m_modelInfo = LLModelInfo();
+        emit modelLoadingError(QString("Could not find file for model %1").arg(modelName));
     }
 
     if (m_modelInfo.model) {
@@ -342,8 +347,7 @@ void ChatLLM::setModelName(const QString &modelName)
 
 void ChatLLM::modelNameChangeRequested(const QString &modelName)
 {
-    if (!loadModel(modelName))
-        qWarning() << "ERROR: Could not load model" << modelName;
+    loadModel(modelName);
 }
 
 bool ChatLLM::handlePrompt(int32_t token)
diff --git a/gpt4all-chat/main.qml b/gpt4all-chat/main.qml
index f72adaae..f8493528 100644
--- a/gpt4all-chat/main.qml
+++ b/gpt4all-chat/main.qml
@@ -62,6 +62,10 @@ Window {
             if (Network.isActive && !currentChat.responseInProgress)
                 Network.sendConversation(currentChat.id, getConversationJson());
         }
+        function onModelLoadingErrorChanged() {
+            if (currentChat.modelLoadingError !== "")
+                modelLoadingErrorPopup.open()
+        }
     }
 
     function startupDialogs() {
@@ -116,6 +120,13 @@ Window {
         Accessible.name: title
     }
 
+    PopupDialog {
+        id: modelLoadingErrorPopup
+        anchors.centerIn: parent
+        shouldTimeOut: false
+        text: currentChat.modelLoadingError
+    }
+
     Rectangle {
         id: header
         anchors.left: parent.left
@@ -126,7 +137,7 @@ Window {
         Item {
             anchors.centerIn: parent
             height: childrenRect.height
-            visible: currentChat.isModelLoaded || currentChat.isServer
+            visible: currentChat.isModelLoaded || currentChat.modelLoadingError !== "" || currentChat.isServer
 
             Label {
                 id: modelLabel
@@ -150,6 +161,31 @@ Window {
                 anchors.horizontalCenterOffset: window.width >= 950 ? 0 : Math.max(-((950 - window.width) / 2), -99.5)
                 enabled: !currentChat.isServer
                 model: currentChat.modelList
+                contentItem: Text {
+                    anchors.horizontalCenter: parent.horizontalCenter
+                    leftPadding: 10
+                    rightPadding: 20
+                    text: currentChat.modelLoadingError !== "" ? "Model loading error..." : comboBox.displayText
+                    font: comboBox.font
+                    color: theme.textColor
+                    verticalAlignment: Text.AlignVCenter
+                    horizontalAlignment: Text.AlignHCenter
+                    elide: Text.ElideRight
+                }
+                delegate: ItemDelegate {
+                    width: comboBox.width
+                    contentItem: Text {
+                        text: modelData
+                        color: theme.textColor
+                        font: comboBox.font
+                        elide: Text.ElideRight
+                        verticalAlignment: Text.AlignVCenter
+                    }
+                    background: Rectangle {
+                        color: highlighted ? theme.backgroundLight : theme.backgroundDark
+                    }
+                    highlighted: comboBox.highlightedIndex === index
+                }
                 Accessible.role: Accessible.ComboBox
                 Accessible.name: qsTr("ComboBox for displaying/picking the current model")
                 Accessible.description: qsTr("Use this for picking the current model to use; the first item is the current model")
@@ -163,7 +199,7 @@ Window {
 
         Item {
             anchors.centerIn: parent
-            visible: !currentChat.isModelLoaded && !currentChat.isServer
+            visible: !currentChat.isModelLoaded && currentChat.modelLoadingError === "" && !currentChat.isServer
             width: childrenRect.width
             height: childrenRect.height
             Row {