diff --git a/src/hooks/ai-generation/graphComposer.test.ts b/src/hooks/ai-generation/graphComposer.test.ts
index bf98655a..0de96702 100644
--- a/src/hooks/ai-generation/graphComposer.test.ts
+++ b/src/hooks/ai-generation/graphComposer.test.ts
@@ -84,9 +84,32 @@ describe('graphComposer', () => {
     expect(warnSpy).toHaveBeenCalled();
   });
 
-  it('maps unknown errors to a stable message', () => {
-    expect(toErrorMessage(new Error('boom'))).toBe('boom');
-    expect(toErrorMessage('bad')).toBe('bad');
-    expect(toErrorMessage('')).toBe('An unexpected error occurred. Please try again.');
+  it('maps a raw model-not-found response to a user-friendly message', () => {
+    const rawError =
+      'models/gemini-3-pro is not found for API version v1beta, or is not supported for generateContent. Service: generativelanguage.googleapis.com (404)';
+    const message = toErrorMessage(new Error(rawError));
+
+    expect(message).toBe(
+      'The selected AI model is unavailable or no longer supported. Please select another model and try again.'
+    );
+    expect(message).not.toContain('generativelanguage.googleapis.com');
+  });
+
+  it.each([
+    ['INVALID API KEY', 'Invalid or missing API key. Please check your AI settings.'],
+    ['Request failed with status 403', 'Access forbidden. Please check your API key permissions.'],
+    ['Quota exceeded', 'Rate limit exceeded. Please wait a moment and try again.'],
+    ['Provider returned 503', 'AI provider server error. Please try again later.'],
+    ['Failed to fetch', 'Network error. Please check your internet connection.'],
+  ])('maps %s without exposing provider details', (rawError, expectedMessage) => {
+    expect(toErrorMessage(new Error(rawError))).toBe(expectedMessage);
+  });
+
+  it('does not expose unknown error details', () => {
+    const expectedMessage = 'The AI request failed. Please try again or choose a different model.';
+
+    expect(toErrorMessage(new Error('upstream service secret detail'))).toBe(expectedMessage);
+    expect(toErrorMessage('raw provider response')).toBe(expectedMessage);
+    expect(toErrorMessage('')).toBe(expectedMessage);
   });
 });
diff --git a/src/hooks/ai-generation/graphComposer.ts b/src/hooks/ai-generation/graphComposer.ts
index c1afc5c5..05cc04ad 100644
--- a/src/hooks/ai-generation/graphComposer.ts
+++ b/src/hooks/ai-generation/graphComposer.ts
@@ -122,46 +122,56 @@ export function toFinalEdges(
 export function toErrorMessage(error: unknown): string {
   if (error instanceof Error) {
     const message = error.message;
+    const normalizedMessage = message.toLowerCase();
 
     if (
-      message.includes('401') ||
-      message.includes('invalid api key') ||
-      message.includes('API key')
+      normalizedMessage.includes('model') &&
+      (normalizedMessage.includes('404') ||
+        normalizedMessage.includes('not found') ||
+        normalizedMessage.includes('unavailable') ||
+        normalizedMessage.includes('not supported'))
+    ) {
+      return 'The selected AI model is unavailable or no longer supported. Please select another model and try again.';
+    }
+
+    if (
+      normalizedMessage.includes('401') ||
+      normalizedMessage.includes('invalid api key') ||
+      normalizedMessage.includes('api key')
     ) {
       return 'Invalid or missing API key. Please check your AI settings.';
     }
-    if (message.includes('403') || message.includes('Forbidden')) {
+    if (normalizedMessage.includes('403') || normalizedMessage.includes('forbidden')) {
       return 'Access forbidden. Please check your API key permissions.';
     }
-    if (message.includes('429') || message.includes('rate limit') || message.includes('quota')) {
+    if (
+      normalizedMessage.includes('429') ||
+      normalizedMessage.includes('rate limit') ||
+      normalizedMessage.includes('quota')
+    ) {
       return 'Rate limit exceeded. Please wait a moment and try again.';
     }
-    if (message.includes('500') || message.includes('internal server error')) {
+    if (
+      /\b5\d{2}\b/.test(normalizedMessage) ||
+      normalizedMessage.includes('internal server error')
+    ) {
       return 'AI provider server error. Please try again later.';
     }
     if (
-      message.includes('network') ||
-      message.includes('fetch') ||
-      message.includes('Failed to fetch')
+      normalizedMessage.includes('network') ||
+      normalizedMessage.includes('fetch')
     ) {
       return 'Network error. Please check your internet connection.';
     }
-    if (message.includes('timeout') || message.includes('timed out')) {
+    if (normalizedMessage.includes('timeout') || normalizedMessage.includes('timed out')) {
       return 'Request timed out. Please try again.';
     }
-    if (message.includes('cursor') || message.includes('parse')) {
+    if (normalizedMessage.includes('cursor') || normalizedMessage.includes('parse')) {
       return 'Failed to parse AI response. Please try a different prompt.';
     }
-    if (message.includes('context length') || message.includes('token')) {
+    if (normalizedMessage.includes('context length') || normalizedMessage.includes('token')) {
       return 'Prompt too long. Try simplifying your request.';
     }
-
-    if (message) {
-      return message;
-    }
-  }
-  if (typeof error === 'string' && error) {
-    return error;
   }
-  return 'An unexpected error occurred. Please try again.';
+  return 'The AI request failed. Please try again or choose a different model.';
 }