Reasoning Gallery One brief · 24 interpretations

One brief · 24 live scenes

One road. Twenty-four ways to draw it.

Luna, Terra, and Sol answered the same canvas brief across their available reasoning levels. Follow one model through every result, match one effort across GPT-5.6, or build your own three-way comparison.

Luna · Terra · Sol17 GPT-5.6 scenes7 reference scenes

Three ways to compare

Start with what you want to hold still.

Keep the model fixed, keep the effort fixed, or choose the exact results yourself.

Experiment overview

What exists, and what was controlled.

The coverage map separates verified shared-effort runs from the standalone Qwen result.

Shared input

The same brief for every result

Write a single HTML file with a full-page canvas and no libraries. Simulate a realistic side-view of a moving car as the main subject. Keep the car visible in the foreground while the background landscape scrolls continuously to create the feeling that the car is driving forward.

Use layered scenery for depth, animate the wheels realistically, add subtle body motion, and create a calm, immersive, seamless loop with cohesive cinematic lighting.

Coverage note

Coverage varies by family. Only verified effort settings are placed on the shared axis; standalone runs remain clearly labeled.

Custom comparison

Choose up to three results

The main slider controls only Luna, Terra, and Sol. Each reference family has its own available settings below.

0 of 3 selected

Model

Result