Claude Opus 5.5
PassTest conditions
Single attempt per test, no retries and no edits afterwards. Run inside a Claude Code agent session, which includes its own system prompt, rather than the raw API.
Exhibit No. 001 · Animation
The most famous vibe check for new models, in its animated form: the pelican has to actually pedal.
> Create a self-contained HTML file with a 2D SVG animation of a pelican riding a bicycleDrawing in code without seeing the result, plus motion that makes physical sense: wheels that roll, pedals that turn, legs that follow them.
Verdicts are our own assessment of each answer.
Single attempt per test, no retries and no edits afterwards. Run inside a Claude Code agent session, which includes its own system prompt, rather than the raw API.
Paste the prompt into the model or agent you want to test. Save the answer as a .html file and open it in a browser.
Animation · Flavio Adamo, 2025
The physics test that went viral in early 2025: a ball with gravity and friction bouncing inside a rotating hexagon.
Animation · AI/ML API model comparison, 2026
A star swells, collapses and explodes, leaving a glowing remnant and a spinning pulsar.