diff --git a/AGENTS.md b/AGENTS.md
index f651ca486..a9f105095 100644
--- a/AGENTS.md
+++ b/AGENTS.md
@@ -1014,6 +1014,14 @@ task push-image    # Build and push multi-platform
 
 5. **Telemetry adds overhead**: Disable with `TELEMETRY_ENABLED=false` for benchmarking
 
+### TUI General Guidelines
+
+- Never use commands to send messages when you can directly mutate children or state.
+- Keep things simple; do not overcomplicate.
+- Create files if needed to separate logic; do not nest models.
+- Never do IO or expensive work in `Update`; always use a `tea.Cmd`.
+- Never change the model state inside of a command use messages and than update the state in the main loop
+
 ## Quick Reference: Key Files
 
 | File | Purpose |
diff --git a/pkg/tui/components/sidebar/sidebar.go b/pkg/tui/components/sidebar/sidebar.go
index 72f1b0617..8f248a220 100644
--- a/pkg/tui/components/sidebar/sidebar.go
+++ b/pkg/tui/components/sidebar/sidebar.go
@@ -47,7 +47,7 @@ type Model interface {
 	SetTokenUsage(event *runtime.TokenUsageEvent)
 	SetTodos(result *tools.ToolCallResult) error
 	SetMode(mode Mode)
-	SetAgentInfo(agentName, model, description string)
+	SetAgentInfo(agentName, model, description string) tea.Cmd
 	SetTeamInfo(availableAgents []runtime.AgentDetails)
 	SetAgentSwitching(switching bool)
 	SetToolsetInfo(availableTools int, loading bool)
@@ -93,6 +93,8 @@ type Model interface {
 	ScrollByWheel(delta int)
 	// IsScrollbarDragging returns true when the scrollbar thumb is being dragged.
 	IsScrollbarDragging() bool
+	// Cleanup cancels any in-flight async operations.
+	Cleanup()
 }
 
 // ragIndexingState tracks per-strategy indexing progress
@@ -143,6 +145,8 @@ type model struct {
 	titleInput         textinput.Model
 	lastTitleClickTime time.Time // for double-click detection on title
 
+	cancelReasoningCheck context.CancelFunc // cancels the in-flight ModelSupportsReasoning call
+
 	// Render cache to avoid re-rendering sections on every frame during scroll
 	cachedLines          []string // Cached rendered lines
 	cachedWidth          int      // Width used for cached render
@@ -248,14 +252,31 @@ func (m *model) SetTodos(result *tools.ToolCallResult) error {
 	return m.todoComp.SetTodos(result)
 }
 
+// reasoningSupportResultMsg carries the async result of a ModelSupportsReasoning check.
+type reasoningSupportResultMsg struct {
+	modelID   string
+	supported bool
+}
+
+// checkReasoningSupportCmd returns a tea.Cmd that checks reasoning support asynchronously.
+func checkReasoningSupportCmd(ctx context.Context, modelID string) tea.Cmd {
+	return func() tea.Msg {
+		supported := modelsdev.ModelSupportsReasoning(ctx, modelID)
+		return reasoningSupportResultMsg{modelID: modelID, supported: supported}
+	}
+}
+
 // SetAgentInfo sets the current agent information and updates the model in availableAgents
-func (m *model) SetAgentInfo(agentName, modelID, description string) {
+func (m *model) SetAgentInfo(agentName, modelID, description string) tea.Cmd {
 	m.currentAgent = agentName
 	m.agentModel = modelID
 	m.agentDescription = description
-	// TODO: this can block for up to 30s on the first call if the cache is cold,
-	// which freezes the TUI. Move to an async command.
-	m.reasoningSupported = modelsdev.ModelSupportsReasoning(context.TODO(), modelID)
+
+	if m.cancelReasoningCheck != nil {
+		m.cancelReasoningCheck()
+	}
+	ctx, cancel := context.WithCancel(context.Background())
+	m.cancelReasoningCheck = cancel
 
 	// Update the provider and model in availableAgents for the current agent.
 	// This is important when fallback models from different providers are used.
@@ -274,6 +295,7 @@ func (m *model) SetAgentInfo(agentName, modelID, description string) {
 		}
 	}
 	m.invalidateCache()
+	return checkReasoningSupportCmd(ctx, modelID)
 }
 
 // SetTeamInfo sets the available agents in the team
@@ -596,9 +618,15 @@ func (m *model) Update(msg tea.Msg) (layout.Model, tea.Cmd) {
 		m.invalidateCache()
 		m.stopSpinner() // Will only stop if no other state needs it
 		return m, nil
-	case *runtime.AgentInfoEvent:
-		m.SetAgentInfo(msg.AgentName, msg.Model, msg.Description)
+	case reasoningSupportResultMsg:
+		if msg.modelID == m.agentModel {
+			m.reasoningSupported = msg.supported
+			m.invalidateCache()
+		}
 		return m, nil
+	case *runtime.AgentInfoEvent:
+		cmd := m.SetAgentInfo(msg.AgentName, msg.Model, msg.Description)
+		return m, cmd
 	case *runtime.TeamInfoEvent:
 		m.SetTeamInfo(msg.AvailableAgents)
 		return m, nil
@@ -1348,3 +1376,11 @@ func (m *model) UpdateTitleInput(msg tea.Msg) tea.Cmd {
 	m.invalidateCache() // Input changes affect rendering
 	return cmd
 }
+
+// Cleanup cancels any in-flight async operations.
+func (m *model) Cleanup() {
+	if m.cancelReasoningCheck != nil {
+		m.cancelReasoningCheck()
+		m.cancelReasoningCheck = nil
+	}
+}
diff --git a/pkg/tui/handlers.go b/pkg/tui/handlers.go
index 515688921..a3377281f 100644
--- a/pkg/tui/handlers.go
+++ b/pkg/tui/handlers.go
@@ -328,11 +328,21 @@ func (a *appModel) handleToggleYolo() (tea.Model, tea.Cmd) {
 }
 
 func (a *appModel) handleToggleThinking() (tea.Model, tea.Cmd) {
-	// Check if the current model supports reasoning
+	if a.cancelThinkingCheck != nil {
+		a.cancelThinkingCheck()
+	}
+	ctx, cancel := context.WithCancel(context.Background())
+	a.cancelThinkingCheck = cancel
+
 	currentModel := a.application.CurrentAgentModel()
-	// TODO: this can block for up to 30s on the first call if the cache is cold,
-	// which freezes the TUI. Move to an async command.
-	if !modelsdev.ModelSupportsReasoning(context.TODO(), currentModel) {
+	return a, func() tea.Msg {
+		supported := modelsdev.ModelSupportsReasoning(ctx, currentModel)
+		return messages.ToggleThinkingResultMsg{Supported: supported}
+	}
+}
+
+func (a *appModel) handleToggleThinkingResult(msg messages.ToggleThinkingResultMsg) (tea.Model, tea.Cmd) {
+	if !msg.Supported {
 		return a, notification.InfoCmd("Thinking/reasoning is not supported for the current model")
 	}
 
@@ -340,20 +350,19 @@ func (a *appModel) handleToggleThinking() (tea.Model, tea.Cmd) {
 	sess.Thinking = !sess.Thinking
 	a.sessionState.SetThinking(sess.Thinking)
 
-	// Persist the change to the database immediately
 	if store := a.application.SessionStore(); store != nil {
 		if err := store.UpdateSession(context.Background(), sess); err != nil {
 			return a, notification.ErrorCmd(fmt.Sprintf("Failed to save session: %v", err))
 		}
 	}
 
-	var msg string
+	var infoMsg string
 	if sess.Thinking {
-		msg = "Thinking/reasoning enabled for this session"
+		infoMsg = "Thinking/reasoning enabled for this session"
 	} else {
-		msg = "Thinking/reasoning disabled for this session"
+		infoMsg = "Thinking/reasoning disabled for this session"
 	}
-	return a, notification.InfoCmd(msg)
+	return a, notification.InfoCmd(infoMsg)
 }
 
 func (a *appModel) handleToggleHideToolResults() (tea.Model, tea.Cmd) {
diff --git a/pkg/tui/messages/toggle.go b/pkg/tui/messages/toggle.go
index cd6802025..99a5a0408 100644
--- a/pkg/tui/messages/toggle.go
+++ b/pkg/tui/messages/toggle.go
@@ -8,6 +8,12 @@ type (
 	// ToggleThinkingMsg toggles extended thinking mode.
 	ToggleThinkingMsg struct{}
 
+	// ToggleThinkingResultMsg carries the async result of reasoning support check.
+	// If Supported is true, thinking is toggled; otherwise a notification is shown.
+	ToggleThinkingResultMsg struct {
+		Supported bool
+	}
+
 	// ToggleHideToolResultsMsg toggles hiding of tool results.
 	ToggleHideToolResultsMsg struct{}
 
diff --git a/pkg/tui/page/chat/chat.go b/pkg/tui/page/chat/chat.go
index 34b77fcae..20e9db3fb 100644
--- a/pkg/tui/page/chat/chat.go
+++ b/pkg/tui/page/chat/chat.go
@@ -1171,6 +1171,7 @@ func (p *chatPage) CompactSession(additionalPrompt string) tea.Cmd {
 
 func (p *chatPage) Cleanup() {
 	p.stopProgressBar()
+	p.sidebar.Cleanup()
 	p.editor.Cleanup()
 }
 
diff --git a/pkg/tui/page/chat/runtime_events.go b/pkg/tui/page/chat/runtime_events.go
index 28071cb31..cb74f1ca9 100644
--- a/pkg/tui/page/chat/runtime_events.go
+++ b/pkg/tui/page/chat/runtime_events.go
@@ -58,10 +58,10 @@ func (p *chatPage) handleRuntimeEvent(msg tea.Msg) (bool, tea.Cmd) {
 
 	case *runtime.ModelFallbackEvent:
 		// Update sidebar with the fallback model immediately so it reflects the switch
-		p.sidebar.SetAgentInfo(msg.AgentName, msg.FallbackModel, "")
+		sidebarCmd := p.sidebar.SetAgentInfo(msg.AgentName, msg.FallbackModel, "")
 		// Notify user when switching to a fallback model, include the reason
 		fallbackMsg := fmt.Sprintf("Model %s failed (%s), switching to %s", msg.FailedModel, msg.Reason, msg.FallbackModel)
-		return true, notification.WarningCmd(fallbackMsg)
+		return true, tea.Batch(sidebarCmd, notification.WarningCmd(fallbackMsg))
 
 	// ===== Stream Lifecycle Events =====
 	case *runtime.StreamStartedEvent:
@@ -102,9 +102,9 @@ func (p *chatPage) handleRuntimeEvent(msg tea.Msg) (bool, tea.Cmd) {
 		return true, nil
 
 	case *runtime.AgentInfoEvent:
-		p.sidebar.SetAgentInfo(msg.AgentName, msg.Model, msg.Description)
+		sidebarCmd := p.sidebar.SetAgentInfo(msg.AgentName, msg.Model, msg.Description)
 		p.messages.AddWelcomeMessage(msg.WelcomeMessage)
-		return true, nil
+		return true, sidebarCmd
 
 	case *runtime.TeamInfoEvent:
 		p.sidebar.SetTeamInfo(msg.AvailableAgents)
diff --git a/pkg/tui/tui.go b/pkg/tui/tui.go
index ced3b6566..72a9ff480 100644
--- a/pkg/tui/tui.go
+++ b/pkg/tui/tui.go
@@ -49,6 +49,8 @@ type appModel struct {
 
 	transcriber *transcribe.Transcriber
 
+	cancelThinkingCheck context.CancelFunc // cancels the in-flight thinking toggle check
+
 	// External event subscriptions (Elm Architecture pattern)
 	themeWatcher      *styles.ThemeWatcher
 	themeSubscription *subscription.ChannelSubscription[string] // Listens for theme file changes
@@ -160,7 +162,7 @@ func New(ctx context.Context, a *app.App) tea.Model {
 	// Make sure to stop the progress bar and theme watcher when the app quits abruptly.
 	go func() {
 		<-ctx.Done()
-		t.chatPage.Cleanup()
+		t.cleanup()
 		t.themeWatcher.Stop()
 	}()
 
@@ -168,6 +170,14 @@ func New(ctx context.Context, a *app.App) tea.Model {
 }
 
 // Init initializes the application
+func (a *appModel) cleanup() {
+	if a.cancelThinkingCheck != nil {
+		a.cancelThinkingCheck()
+		a.cancelThinkingCheck = nil
+	}
+	a.chatPage.Cleanup()
+}
+
 func (a *appModel) Init() tea.Cmd {
 	cmds := []tea.Cmd{
 		a.dialog.Init(),
@@ -350,7 +360,7 @@ func (a *appModel) Update(msg tea.Msg) (tea.Model, tea.Cmd) {
 
 	case messages.ExitSessionMsg:
 		// /exit command exits immediately without confirmation
-		a.chatPage.Cleanup()
+		a.cleanup()
 		return a, tea.Quit
 
 	case messages.NewSessionMsg:
@@ -407,6 +417,9 @@ func (a *appModel) Update(msg tea.Msg) (tea.Model, tea.Cmd) {
 	case messages.ToggleThinkingMsg:
 		return a.handleToggleThinking()
 
+	case messages.ToggleThinkingResultMsg:
+		return a.handleToggleThinkingResult(msg)
+
 	case messages.ToggleHideToolResultsMsg:
 		return a.handleToggleHideToolResults()
 
@@ -525,11 +538,11 @@ func (a *appModel) Update(msg tea.Msg) (tea.Model, tea.Cmd) {
 		return a, nil
 
 	case dialog.ExitConfirmedMsg:
-		a.chatPage.Cleanup()
+		a.cleanup()
 		return a, tea.Quit
 
 	case messages.ExitAfterFirstResponseMsg:
-		a.chatPage.Cleanup()
+		a.cleanup()
 		return a, tea.Quit
 
 	case chat.EditorHeightChangedMsg: