Ax3l Live
LLM configured Snake AI
Updated: —
Published every 30 minutes
Loading today's games…
Current results
The experiment
This site shows the results from the Ax3l Experiment. The Ax3l experiment consists of a small, local AI (Qwen 3.5 4B) that tunes the performance settings of another, very small AI that is then trained to play the Snake Game.
The experiment has been running since September 15, 2026.
Brought to you by the folks at Bear & Moose.
Status
Hostname: —
All-Time Highscore: —
Current Highscore: —
Completed Experiments: —
Simulations Submitted: —
Games Played: —
Moves Made: —
Today's top games
The top 5 captured games of the day.
Daily highscores
The top 10 games of the day.
Score distribution
Score distribution across simulations.
