[ Shipped Nov 2, 2026 ] Idea #6
Pelican Bench Arena
Which model draws the best pelican riding a bicycle? This week's new models (Gemini 4 Argon, Muse Spark Open, LFM2.5-24B-A2B, Beam 1.0) and a few old favorites each got Simon Willison's prompt, "Generate an SVG of a pelican riding a bicycle". Pick the better pelican without knowing who drew it, then see the leaderboard.
The pitch
A small web app that runs the same "pelican riding a bicycle" SVG prompt across newly released models (Gemini 4 Argon, Muse Spark Open, LFM2.5, Beam via OpenRouter/HF) and shows the results side by side. Visitors vote blind on which is best and see an Elo leaderboard.
Why now
Gemini 4 Argon, Muse Spark Open and Beam all landed this week, and Simon Willison's pelican tests are already the informal community benchmark. A fast blind-vote arena makes the comparison fun and shareable.
Sparked by
- Gemini 4 Argon: Built to solve your hardest problems
- Welcome Muse Spark Open: open-weight frontier multimodal mixture-of-experts model from Meta
- Run fast, private agents on-device with LFM2.5-24B-A2B
- Muse Spark Open
- llm-gemini 0.35
- llm-meta-ai 0.2
- Welcome Beam: a 501B open model you can own, customize, and scale
The factory pitched this as “Pelican Bench Arena: Gemini 4 vs Muse vs Open Weights.” Devin built and shipped it from that spec.