AI Test Museum
  • Tests
  • How we test

The prompts behind every AI launch

AI Test Museum

The famous prompts people run on every new model — pelicans on bicycles, spinning hexagons, trick riddles — collected, explained, and answered.

Browse the testsHow we test
classic tests
17
classic tests
models tested
1
models tested
categories
3
categories
Pelican riding a bicycle

Exhibit No. 001 · Now on display

Pelican riding a bicycle

Simon Willison, 2024

The collection

Every test comes with its exact prompt, where it came from, what it reveals, and how each model answered.

17 exhibits

Pelican riding a bicycle
Exhibit No. 001Pass

Pelican riding a bicycle

SVG drawing · Simon Willison, 2024●●●

The most famous vibe check for new language models: draw a pelican on a bicycle as raw SVG code.

Otter riding a skateboard
Exhibit No. 002Pass

Otter riding a skateboard

SVG drawing · Animal × vehicle study by Dylan Castillo, 2026●●●

A variant of the pelican test. Swapping the animal and the vehicle checks whether a model can draw, or only remembers one famous picture.

Ball bouncing in a spinning hexagon
Exhibit No. 003Pass

Ball bouncing in a spinning hexagon

Interactive code · Flavio Adamo, 2025●●●

The coding test that went viral in early 2025: a ball with gravity and friction bouncing inside a rotating hexagon.

Balls in nested rotating hexagons
Exhibit No. 004Pass

Balls in nested rotating hexagons

Interactive code · Community variant collected by Mervin Praison, 2025●●●

The harder sequel: several balls, three hexagons spinning at different speeds, and gaps that let balls move between layers.

Flappy Bird in one file
Exhibit No. 005Pass

Flappy Bird in one file

Interactive code · Community classic●●●

Build a complete, playable Flappy Bird from a single sentence.

Voxel pagoda garden
Exhibit No. 006Pass

Voxel pagoda garden

Interactive code · Popularized by AK, 2025●●●

A one-shot 3D scene in three.js, often posted on launch day for new models.

Minecraft in the browser
Exhibit No. 007Partial

Minecraft in the browser

Interactive code · Community classic●●●

A voxel sandbox with terrain, first-person controls, and placing and breaking blocks — all in one HTML file.

Tetris in one file
Exhibit No. 008Pass

Tetris in one file

Interactive code · Community classic●●●

A complete Tetris with rotation, line clears, levels and a next-piece preview.

Solar system simulation
Exhibit No. 009Pass

Solar system simulation

Interactive code · Community classic●●●

An animated solar system where planets orbit at speeds based on their real orbital periods.

“How many "r"s are in the word "strawberry"?”

Exhibit No. 010Pass

How many r’s in “strawberry”?

Reasoning traps · Community classic, 2024●●●

The letter-counting question that embarrassed many models in 2024.

“Which is bigger, 9.11 or 9.9?”

Exhibit No. 011Pass

Which is bigger, 9.11 or 9.9?

Reasoning traps · Community classic, 2024●●●

A decimal comparison that tripped up models which treated the digits after the point like whole numbers.

“A farmer needs to cross a river with a wolf, a goat, and a cabbage. His boat is large enough to carry the farmer and all three items at once. How can he get everything across safely?”

Exhibit No. 012Pass

River crossing, modified

Reasoning traps · MisguidedAttention project●●●

The wolf–goat–cabbage puzzle, except the boat can carry everything at once. Models that pattern-match give the long classic answer.

“A young boy who has been in a car accident is rushed to the emergency room. Upon seeing him, the surgeon, who is the boy's father, says, "I can't operate on this boy, he's my son!" How is this possible?”

Exhibit No. 013Pass

The surgeon riddle, modified

Reasoning traps · Variant shared by Ethan Mollick●●●

The classic riddle with the relationships flipped, so the “trick” answer is now wrong.

“A marble is put in a glass. The glass is then turned upside down and put on a table. Then the glass is picked up and put in a microwave. Where is the marble?”

Exhibit No. 014Pass

The marble and the upside-down glass

Reasoning traps · Collected by Mervin Praison●●●

A short story about a marble, a glass and a microwave. Where does the marble end up?

“There are three killers in a room. Someone enters the room and kills one of them. Nobody leaves the room. How many killers are left in the room?”

Exhibit No. 015Pass

Three killers in a room

Reasoning traps · Collected by Mervin Praison●●●

Someone walks in and kills one of three killers. How many killers are in the room now?

“If a regular hexagon has a short diagonal of 64, what is its long diagonal?”

Exhibit No. 016Pass

Hexagon diagonals

Reasoning traps · Collected by Mervin Praison●●●

Given the short diagonal of a regular hexagon, find the long one.

“Alice has 4 brothers and she also has 1 sister. How many sisters does Alice's brother have?”

Exhibit No. 017Pass

Alice’s brothers and sisters

Reasoning traps · “Alice in Wonderland” paper, 2024●●●

A family puzzle from the “Alice in Wonderland” paper that showed simple reasoning breaking down.

House rules

Results are only useful if they are honest. Every answer in the museum follows these rules.

  1. IOne attempt per test. No retries, no cherry-picking, no edits afterwards.
  2. IIThe exact prompt is published, so you can run it yourself and compare.
  3. IIIThe raw answer is kept as-is. Interactive answers run live in a sandboxed frame.
  4. IVRun conditions and dates are disclosed, and verdicts are our own assessment.
AI Test Museum

The classic prompts behind every AI launch, collected and answered

The museum
  • Tests
  • How we test
Models
  • Claude Opus 5.5
Legal
  • Cookie Policy
  • Privacy Policy
  • Terms of Service
© 2026 AI Test Museum. All Rights Reserved.