Observed arrival · 2026-11-21
Robo Code Bench puts AI models on a VEX field
A research project tests whether three AI models can write programs to navigate a real VEX AIM robot through four games.
○Open
⊠Login
$Paid
†Ads
✦Pretty
●Pro
◎Niche
◉Human
⚑Risk
ƒJS
- For
- AI and robotics benchmark researchers
- Worth noticing
- The paper flags battery level, wheel slip, and ball movement as sources of noise, and does not claim statistical significance.
Field notes
The harness gives each model the same kinds of inputs: game rules, the VEX AIM Python API reference, and an overhead photo with a ruler for scale. A model returns a complete Python program, and the page describes recording camera frames, console logs, and telemetry across runs. The authors caution that each model played each game once; battery level, wheel slip, and how the ball rolls can affect results.
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