Skip to content
TheChatArchiveThe Chat Archive
o4-mini-2025-04-16· OpenAIHugging Face LMArena Human Preference 140K

Generate a picture, a cat who is playing chess with a dragon!

From a Chatbot Arena battleThe same person sent this prompt to two anonymous models at once. Compare it with Claudeclaude-opus-4-20250514’s answer.
  1. User

    Generate a picture, a cat who is playing chess with a dragon!

  2. Assistant

    Here’s a two-part solution—

    1. A ready-to-use prompt for your favorite AI image generator (DALL·E, Midjourney, Stable Diffusion, etc.), plus
    2. A quick ASCII sketch to spark your imagination!

    1) AI-Art Prompt

    “A whimsical, cozy scene: on a round wooden table draped with a velvet cloth sits a carved mahogany chessboard. 
    On one side, a clever tabby cat wearing a tiny spectacles perches on a plush cushion, paw poised over a pawn. 
    Across from it, a slender emerald-green dragon with gold-flecked scales leans in, smoke curling from its nostrils 
    as it contemplates its next move. Soft lantern light casts warm shadows; background hints at ancient stone walls 
    and shelves of dusty tomes. Highly detailed, painterly style, vibrant but moody color palette, cinematic lighting.”
    

    Feel free to tweak:

    • Change “tabby” to “black-and-white kitty” or “fluffy Maine Coon.”
    • Swap “emerald-green” for “ruby-red” or “ice-blue.”
    • Adjust lighting: “sunset glow” / “moonlit” / “candlelit.”

    2) Mini ASCII Sketch

           __/\\__  
          / o  o \      _____
         (    ^   )    /     \__
          \ \__/ /     \  D  /  \       /\_/\
           \____/  ____  \__/  /      =( °w° )=
            /  \  /    \______/        )   (  //
           /____\/                      (__ __)//
       Cat                   Chessboard                 Dragon
    

    (Imagine the cat on the left, the dragon’s head peeking on the right, chan‍ging a chess move in the center!)


    Have fun! Let me know if you’d like a different art style, mood, or more visual details. 🐱🐉🎲

Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.