Skip to content
TheChatArchiveThe Chat Archive
Claudeclaude-opus-4-20250514· AnthropicHugging Face LMArena Human Preference 140K

Generate a few cool and interesting prompts to test the capabilities of an image…

From a Chatbot Arena battleThe same person sent this prompt to two anonymous models at once. Compare it with Claudeclaude-3-7-sonnet-20250219’s answer.
  1. User

    Generate a few cool and interesting prompts to test the capabilities of an image generation model

  2. Claude Assistant

    Here are some creative prompts to test different aspects of an image generation model:

    Technical Capability Tests:

    "A transparent glass sculpture of a dragon filled with swirling galaxy patterns, sitting on a reflective marble pedestal in a minimalist white gallery, dramatic lighting from above"

    "An oil painting in the style of Vermeer depicting a modern programmer working on a laptop by candlelight, with code reflections visible in their glasses"

    Complex Scene Composition:

    "A bustling floating market on Mars, with vendors in space suits selling exotic alien fruits under biodome tents, Earth visible in the rust-colored sky"

    "A macro photograph of a dewdrop on a spider web, with an entire miniature medieval castle reflected perfectly inside the droplet"

    Unusual Perspectives:

    "A bird's eye view of a giant's tea party, where the teacups are repurposed swimming pools and cookies are the size of tennis courts, with tiny humans observing from the garden below"

    "The view from inside a violin looking out through the f-hole during a concert, with dramatic stage lights creating patterns through the opening"

    Texture and Material Challenges:

    "A steampunk butterfly made entirely of brass gears, copper pipes, and stained glass wings, perched on a Victorian pocket watch"

    "A hyperrealistic marble statue that appears to be melting like ice cream on a hot day, with the liquid marble pooling dramatically at the base"

    These prompts test various capabilities including transparency, reflections, style mimicry, complex compositions, unusual perspectives, detailed textures, and physics-defying scenarios.

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