Skip to content
TheChatArchiveThe Chat Archive
DeepSeekdeepseek-r1-0528· DeepSeekHugging 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 Claudeclaude-sonnet-4-20250514’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. DeepSeek Assistant

    Futuristic Image Generation Test Prompt:

    Prompt:
    "Generate a photorealistic scene of a Neo-Tokyo skyline in 2150 at twilight. The foreground features a hover-train gliding silently on a magnetic track, reflecting neon lights from holographic billboards. In the midground, a cybernetic garden with bioluminescent trees and floating drones tending to exotic flora. The background includes glass-domed skyscrapers connected by aerial walkways, with anti-gravity vehicles zipping through smog-laden clouds. A humanoid AI with translucent circuitry skin stands on a transparent observation deck, watching a holographic constellation map projected into the sky. Style: Cinematic lighting, hyper-detailed, volumetric fog, and chromatic aberration."


    Why This Prompt Tests AI Model Capabilities:

    1. Complex Scene Composition
      Tests the model’s ability to layer foreground, midground, and background elements cohesively without visual clutter. Weak models may merge objects or ignore spatial depth.

    2. Futuristic Conceptualization
      Challenges creativity in designing novel elements (e.g., "cybernetic gardens," "translucent circuitry skin"). Models must extrapolate beyond real-world references without guidance.

    3. Material and Lighting Nuance
      Requires rendering:

      • Reflections (hover-train, glass domes),
      • Glowing effects (bioluminescent trees, holograms),
      • Atmospheric depth (volumetric fog, smog).
        Poor models oversimplify textures or lighting.
    4. Dynamic Elements
      Incorporates motion (hover-train, anti-gravity vehicles) and transparency (observation deck, holograms). Tests if AI implies movement realistically.

    5. Style Consistency
      Demands adherence to "cinematic hyper-realism" with photorealistic details (e.g., circuitry skin, drone mechanics) while avoiding cartoonish or surreal artifacts.

    6. Contextual Coherence
      Ensures all elements fit the "Neo-Tokyo 2150" theme. Weak models may insert anachronistic objects or inconsistent architecture.


    Key Comparison Metrics:

    • Detail Accuracy: How precisely are intricate elements (e.g., bioluminescent trees, circuitry skin) rendered?
    • Atmosphere: Does volumetric fog integrate naturally with lighting?
    • Innovation: Are futuristic elements uniquely interpreted or generic?
    • Artifacts: Check for glitches in transparency/reflections or distorted proportions.

    This prompt stresses imagination, technical precision, and compositional balance—revealing gaps in low-quality models through chaotic outputs, missing details, or stylistic inconsistency.

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