Observed arrival · 2026-10-17
Hot dog. Not hot dog.
A playful benchmark compares ten AI models on 1,000 food photos, tracking accuracy, speed, and cost at one very specific classification task.
○Open
⊠Login
$Paid
†Ads
✦Pretty
●Pro
◎Niche
◉Human
⚑Risk
ƒJS
- For
- Developers comparing vision-model cost and accuracy
- Worth noticing
- The benchmark separates missed hot dogs from false positives and reports 95% intervals alongside cost per 1,000 images.
Field notes
The table distinguishes missed hot dogs from false alarms, reports 95% intervals, and exposes individual error photos. A human challenge offers 50 images under a timer; the page also links to a dataset, a reproduction path, and a MotherDuck demo. The displayed results put the cheapest entry at 1.7¢ per 1,000 images and the highest-priced at $10.27.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
○OpenPublic substance visible
✦PrettyNotable craft visible
●ProPolished or operationally mature
◎NicheUnusually specific use
One card from the complete issue