Qwen3 demo

Every scenario DeepQA drove in the browser, in plan order, with the 32 screenshots it captured along the way. A passing scenario is evidence too.

1. S1

   Send a standard chat message

   4 steps, 4 screenshots

   high✓ pass

   

   

   

   

   • Loaded the Qwen3 demo homepage with chat input and model settings visible.
   • Entered prompt 'Hello, can you help me?' into the chat input textbox.
   • Clicked submit button, conversation started and model began streaming its thought process and response.
   • AI response generated successfully with thinking duration (5.35s), total duration (6.27s), and action buttons (Copy, etc.).
   • Navigated to the Qwen3 demo application at https://qwen-qwen3-demo.hf.space/.
   • Entered 'Hello, can you help me?' into the chat input textbox and submitted.
   • Observed live thinking process and streaming generation.
   • Verified AI response completed displaying the answer, thinking duration badge 'End of Thought (5.35s)', total duration '6.27s', and action buttons including 'Copy'.
2. S2

   Start a new conversation and verify history

   4 steps, 3 screenshots

   high✓ pass

   

   

   

   • Sent 'Test message' and received response 'Hello! It looks like you're testing the system. How can I assist you today?' from Qwen3.
   • Clicked 'New Conversation' button; the main view reset to the welcome screen and 'Test message' was retained in the history list.
   • Loaded the demo application with an empty initial state and default welcome screen.
   • Submitted 'Test message' to the model and received a full completion response with thinking time.
   • Clicked 'New Conversation' which reset the chat pane to the welcome screen and preserved 'Test message' in the history list in the sidebar.
3. S3

   Adjust Thinking Budget using slider

   3 steps, 2 screenshots

   high✓ pass

   

   

   • Navigated to Qwen3 demo page where the settings panel is visible showing Thinking Budget slider and spinbutton currently set to 38k.
   • Navigated to https://qwen-qwen3-demo.hf.space/ and located the Settings panel containing the Thinking Budget slider and spinbutton.
   • The default Thinking Budget for model Qwen3-235B-A22B was 38k.
   • Adjusting the value in the spinbutton updated the slider position and tooltip to 20k.
   • Navigating and adjusting the slider using keyboard controls (ArrowLeft/ArrowRight) smoothly updated the slider position and immediately synced the Thinking Budget spinbutton value to 19k and 20k.
4. S4

   Empty message validation

   3 steps, 4 screenshots

   medium✓ pass

   

   

   

   

   • Loaded the Qwen3 demo page with empty chat input.
   • Clicked the Submit button with an empty chat input; the application did not trigger a submission, show loading state, or add an empty chat bubble.
   • Verified that pressing Enter or clicking Submit with an empty input does not trigger any submission or add an empty message bubble.
   • Navigated to https://qwen-qwen3-demo.hf.space/ and ensured the chat input textbox was empty.
   • Clicked the submit button with an empty input and verified that no submission was made, no loading state was triggered, and no empty message bubble was added to the chat.
   • Tested pressing the Enter key on the empty chat input and verified that no submission occurred.
5. S5

   Adjust Thinking Budget using spinbutton

   3 steps, 2 screenshots

   medium✓ pass

   

   

   • Initial page loaded with Settings drawer open showing Thinking Budget slider and spinbutton with value 38.
   • Typed 16 into Thinking Budget spinbutton and confirmed the value was updated to 16.
   • Navigated to the application and verified the Settings drawer contains the Thinking Budget controls.
   • Entered 16 into the Thinking Budget spinbutton, successfully updating the input value to 16.
6. S6

   Slash command suggestions trigger

   3 steps, 2 screenshots

   medium✓ pass

   

   

   • Focused the chat input textbox and entered '/', which successfully triggered suggestion popup menus for 'Make a plan', 'Help me write', 'Start a business', 'Achieve my goals', and 'Successful interview'.
   • Loaded the Qwen3 demo application at https://qwen-qwen3-demo.hf.space/.
   • Focused the chat input textbox with placeholder 'Enter "/" to get suggestions'.
   • Typed '/' into the chat textbox and observed suggestion menus pop up with prompt options ('Make a plan', 'Help me write', 'Start a business', etc.).
   • Confirmed that the slash command suggestion menu works as expected.
7. S7

   Toggle Thinking mode

   2 steps, 3 screenshots

   medium✓ pass

   

   

   

   • Navigated to https://qwen-qwen3-demo.hf.space/ and found the 'Thinking' mode button in the prompt toolbar.
   • Clicked the 'Thinking' mode toggle button in the prompt toolbar, triggering state update in the UI.
   • Navigated to the Qwen3 Gradio interface at https://qwen-qwen3-demo.hf.space/.
   • Located the Thinking mode toggle button in the prompt toolbar alongside the input textbox.
   • Clicked the Thinking mode toggle button, successfully toggling its active state in the interface.
8. S8

   View API documentation modal

   2 steps, 2 screenshots

   medium✓ pass

   

   

   • Clicked 'Use via API' button; the URL updated to '?view=api' and the API documentation modal opened with endpoint details and code snippets.
   • Navigated to the home page and clicked the 'Use via API' button in the footer.
   • The URL parameter '?view=api' was appended properly without reloading the page.
   • The API documentation overlay opened displaying 18 API endpoints, client installation instructions, and code snippets in multiple programming languages.
9. S9

   Submit extremely long message

   3 steps, 3 screenshots

   medium✓ pass

   

   

   

   • Submitted 5000 character input; application accepted it and started streaming the response cleanly without UI overflow or crash.
   • Navigated to https://qwen-qwen3-demo.hf.space/ and entered a 5000-character test message into the chat input textbox.
   • Clicked the Submit button; the application successfully processed the 5000-character message and began generating a response with thinking stream without any crash or visual distortion.
10. S10

    Close API documentation modal

    3 steps, 2 screenshots

    low✓ pass

    

    

    • Navigated to https://qwen-qwen3-demo.hf.space/?view=api and verified the API documentation modal overlay is displayed.
    • Clicked the close button on the API documentation modal; the overlay was dismissed and the core chat interface is visible and intact.
    • The API documentation modal rendered properly when visiting https://qwen-qwen3-demo.hf.space/?view=api.
    • Clicking the modal close button successfully dismissed the modal overlay and returned to the core chat interface without errors or blank pages.
11. S11

    View Gradio settings modal

    2 steps, 2 screenshots

    low✓ pass

    

    

    • Loaded the home page and located the Settings button in the footer.
    • Clicked the Settings footer button; the URL updated to '?view=settings' and the Gradio settings modal opened displaying Language and Progressive Web App sections.
    • Navigated to https://qwen-qwen3-demo.hf.space/ and clicked the Settings button in the footer.
    • The page URL updated to https://qwen-qwen3-demo.hf.space/?view=settings and the general Gradio settings modal opened displaying Language and Progressive Web App settings.
12. S12

    Out of bounds Thinking Budget input

    3 steps, 3 screenshots

    low✓ pass

    

    

    

    • Initial page view loaded with Thinking Budget spinbutton visible with value 38.
    • Entering a negative value (-10 or -50) into the Thinking Budget spinbutton was rejected and clamped to the minimum allowed value of 1.
    • The Thinking Budget spinbutton clamps negative input values (e.g. -10 and -50) to the minimum allowed bound of 1 upon pressing Enter or clicking away.