Observed arrival · 2026-10-15
Tested Locally’s AI arena keeps receipts
A public log of AI models tested on real tasks, including game matches, with runs, costs, and results.
○Open
⊠Login
$Paid
†Ads
✦Pretty
●Pro
◎Niche
◉Human
⚑Risk
ƒJS
- For
- Readers comparing AI models on practical tasks
- Worth noticing
- The site says two runs were left out under its published rules; its current log reports 29 calls and $10.98 spent.
Field notes
The homepage presents the work as both an experiment log and an episodic show: Arena entries include model matchups, while The Café Show is described as turning matches into animated episodes. A separate build log covers tools and pipeline tests, including one listed test costing $0.0000204. The site says its data, charts, and text may be shared unchanged with attribution under CC BY-ND 4.0.
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
ƒJavaScriptBrowser-side code central
One card from the complete issue