Observed arrival · 2026-09-09
Agentic Robotics Benchmark: an exam for autonomous machines
A robotics evaluation platform that runs autonomous agents through sandboxed tasks and scores their resulting controllers, plans, designs, or code on hidden seeds.
○Open
⊠Login
$Paid
†Ads
✦Pretty
●Pro
◎Niche
◉Human
⚑Risk
ƒJS
Field notes
The benchmark separates agent work from evaluation: the ALE engine blocks networking, freezes /home/user/submission, and runs verification only after the agent stage ends. Scores are normalized against a reference implementation's measured anchor, capped, and aggregated with skipped tasks counting as zero. The homepage currently shows drone_hover as a UAV Control task with a three-hour agent budget and mean episode reward as its metric.
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