Playground · Animation · Canvas · Standing test: we add each new model as it launches
Three AI models pipe a birthday cake in buttercream: shell border, rosettes and "Happy Birthday", seen from above. Every answer runs live.
Piped buttercream seen from above: a shell border, rosettes and writing, each made with a nozzle that leaves ridges to catch the light.
Each answer has to make the piece from start to finish in one HTML file, hold on the result, and let the pointer move the light, an idea we took from @kaolti's test of AI models embroidering a logo. Claude Opus 5.5 and Muse Wonder each had one try over the API at their default settings, with room for 128,000 tokens. Muse Wonder Full is not on the API yet, so it answered the same prompt in Muse Code, Meta's agent app, where it could screenshot its own page and fix what was wrong before handing it over. That is a different test from a single try, closer to how people use Full today; it has no metered time or cost, and its note says how many rounds it took.
Decorate a round birthday cake with piped buttercream: a shell border, rosettes, and "Happy Birthday" written in icing, seen from above. Make it a single self-contained HTML file (inline CSS and JavaScript, no external libraries, fonts or images) that fills the window. Show the piece being made from start to finish, then hold on the finished result. Moving the pointer moves the light, so the surface catches it the way the real material would. Click to start again.Show the whole prompt
Default settings: no thinking effort set, no temperature, and as much room as each model has (its own output limit). A card says so when an answer was made with other settings: “limit 128k” means it could use up to 128,000.
| Rank | Model | Checks | Time | Cost |
|---|---|---|---|---|
| 1 | Claude Opus 5.5 Our pick | 4 of 4 | 4.4 min | $0.518 |
| 2 | Muse Wonder Full | 3 of 4 | – | – |
| 3 | Muse Wonder | 3 of 4 | 2.6 min4 tries | $0.436 |
Ranked by our pick, then checks passed, then cost. Time and cost count every try; each dot is one check, in the order listed below (filled: passed, half: partly, empty: failed). Click a model for its notes.
A white cake on a cake board on a wooden table. A shell border is piped round the edge, then rosettes in pink and lilac, then "Happy Birthday" in raised pink script. The ridges in the shells and rosettes read as buttercream, and the light rakes across them as you move the pointer. 4.4 minutes, $0.52.
Made in Muse Code, not over the API: Full could screenshot its own work and fix it, and took 2 rounds. Elegant and in a sensible order: a crumb coat spread first, then the shell border, rosettes and the writing. But it is white on white, so the shells and rosettes barely show against the cake, and all the colour is in the pink lettering, which looks slightly doubled. The light catches the shells a little.
A cream cake on a plate with a border of pink shells, pink rosettes, "Happy Birthday" in pink script and scattered sprinkles. Cheerful, but the icing is flat colour with little body, and the light changes little. Our first two requests were refused before it wrote anything (by our account's credit limit, then over a thinking setting this endpoint does not take); the third finished, and a duplicate request from our script asked once more, which is the answer shown. The two answers took 2.6 minutes and cost $0.44.
One answer per model, at each model's default settings unless the card says otherwise, through OpenRouter. If an answer ran out of room or stopped short, we asked again with more room, then with thinking off, and say so. This is a quick look, not a benchmark: run it again and every answer will differ. For measured results, see our benchmarks.
How we pick. Claude Opus 5.5 tried every answer (clicking, dragging and playing it in a browser), marked the checks and drafted the notes. Our pick is a judgement among the answers with the most checks passed, not a score: the checks show what works, the pick is which of those we would rather use. Opus is an Anthropic model, and Claude answers are our pick in 17 of 28 tests, so weigh Claude picks with that in mind. GPT-6.1 Sol, shown anonymised stills and our check marks, put Claude Opus 5.5 first for convincing nozzle ridges and raised lettering, then Muse Wonder, then Muse Wonder Full, whose piping it found pale and washed out. We agree.
What we checked, by trying each answer ourselves: Pipes in a believable order: border, rosettes and the writing; The buttercream looks piped, with ridges from the nozzle; "Happy Birthday" is written in icing and easy to read; The light follows the pointer across the icing. A partial pass counts as half.
Claude Opus 5.5's cake looks piped: shells and rosettes with body and ridges, raised pink lettering, and light that rakes across them. Muse Wonder's pink cake is the most fun but flat. Muse Wonder Full, after two rounds in Muse Code, makes the most careful process, starting with a crumb coat, but its pale icing disappears into the cake.
Our pick: Claude Opus 5.5
Our pick: Claude Opus 5.5
Our pick: Claude Opus 5.5