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

create a prompt which tests AI model image generation capabilities. Let's do something…

From a Chatbot Arena battleThe same person sent this prompt to two anonymous models at once. Compare it with DeepSeekdeepseek-r1-0528’s answer.
  1. User

    create a prompt which tests AI model image generation capabilities. Let's do something futuristic. Point out why the created prompt is suitable for comparing AI models capabilities.

  2. Claude Assistant

    Here's a comprehensive prompt for testing AI image generation capabilities:

    Test Prompt: "A bustling cyberpunk marketplace in Neo-Tokyo 2087, viewed from a low angle looking up at towering holographic advertisements in multiple languages floating between glass skyscrapers. In the foreground, a chrome-plated android vendor with visible circuitry patterns sells glowing data crystals from a levitating stall. Neon rain falls upward due to gravity manipulation technology, creating luminous streaks against the dark sky. The scene is lit by a mix of cool blue hologram light, warm orange street lamps, and purple lightning from an approaching storm. Include fine details like steam rising from food vendors, reflection distortions in puddles, and people wearing translucent smart-clothing with embedded LED patterns."

    Why this prompt effectively tests AI model capabilities:

    This prompt challenges multiple technical and creative aspects simultaneously. The complex lighting scenario with three distinct light sources tests the model's understanding of how different colored lights interact and cast shadows. The specific architectural elements like "glass skyscrapers" and "levitating stall" evaluate spatial reasoning and physics comprehension.

    The prompt demands high detail rendering through elements like "visible circuitry patterns," "reflection distortions," and "embedded LED patterns," which separates models that can handle fine details from those that produce blurry or simplified outputs. The unusual physics element of "neon rain falling upward" tests creative interpretation and the ability to visualize impossible scenarios convincingly.

    Compositionally, the low-angle perspective and foreground-to-background depth create a complex spatial challenge. The multicultural element of "multiple languages" tests text rendering capabilities, while the specific aesthetic requirements of "cyberpunk" evaluate style consistency and genre understanding.

    The atmospheric details like steam, reflections, and weather effects test the model's ability to create convincing environmental storytelling rather than just placing objects in space. This combination makes it easy to compare which models excel at technical precision versus creative interpretation versus overall coherence.

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

create a prompt which tests AI model image generation… · The Chat Archive