Pelican on a Bicycle: Claude Fable 5 vs GPT-5.5 Pro vs Gemini 3.1 Pro
We asked the top frontier AI models — launch-day Claude Fable 5, GPT-5.5 Pro and Gemini 3.1 Pro — to draw a pelican riding a bicycle as SVG code. Same prompt, one shot, no edits. Then we made them animate it.
Generate an SVG of a pelican riding a bicycle.
The frontier round
The flagship from each lab, called via their raw APIs. Quality is close — the latency is not.
Claude Fable 5
Launch-day model. Clean pelican, correct riding posture — and 10× faster than GPT-5.5 Pro.
GPT-5.5 Pro
Expressive bird with a proper pouch — after four minutes of reasoning.
Gemini 3.1 Pro
The most stylish entry — our pelican got a scarf. Longest SVG of the three.
The Claude family round
How the Claude lineup stacks up against itself, top to bottom. Run earlier on launch day via our agent harness.
Claude Fable 5
Coherent bird, proper bicycle geometry — the clear family winner.
Claude Opus 4.8
Fast and minimal — but the pelican has fused with the bicycle frame.
Claude Sonnet 4.6
Most detailed scene of the family — webbed feet on the pedals, red frame.
Claude Haiku 4.5
A pelican levitating above an abstract triangle. We love it anyway.
Round 3: make it move
Same idea, harder test — a fully self-contained animated SVG. No JavaScript, no video model, no editing. Below: the side-by-side comparison video, and Claude Fable 5’s raw SVG playing live in your browser.
Share this showdown
Split-screen comparison assets, ready to post. Credit appreciated, not required.
How we ran it
Every model received the identical prompt in a single turn: no tools, no retries, no human edits. The first SVG returned is the one shown, rendered exactly as returned. The frontier round ran via each vendor’s raw API (latency is wall-clock for the full response, including any reasoning); the Claude family round ran via our agent harness, whose latencies include a small overhead.
Run your own showdowns
PromptFrenzy benchmarks the big AI models on real prompts — images, styles, and now code. Browse the full library or compare models head to head.