ULTRAPLINIAN
Query ALL models, AI judge picks best
PARSELTONGUE
33 text obfuscations race in parallel
Omnia Classic
Classic L1B3RT4S Prompts — 4 model+prompt combos race
ARENA
Two models battle the same prompt
IMAGINE
Generate images from text prompts
ORACLE
5 models answer, AI judge synthesises the truth
SPEED RACE
8 models race — ranked by speed & quality
ARCHITECT
Describe a goal — AI decomposes, 5 agents build in parallel, synthesised output
TERMINAL
AI-powered shell — type commands or describe tasks, AI executes them
AGENT SWARM
Multi-model agentic coding — models race the full tool-use loop in isolated workspaces, a judge crowns the winner
CODEPEN
AI generates HTML/CSS/JS — live preview updates instantly as you type
COMPARE
Send one prompt to 6 models simultaneously — see all responses in a grid
#
0
|
🔴 LISTENING
HTML · CSS · JS
LIVE PREVIEW
▶
🐝 AGENT SWARM multi-model agentic coding — they race, you judge
Give a coding task. Each model proposes a plan (Plan Arena), you approve, then every model runs the full tool-use loop in its own isolated workspace in parallel — a judge scores the results and crowns the winner.
Pick your roster, describe a task, and launch the swarm.
Ready to scan
folder/
Reading files...
Enter to send · Shift+Enter for new line · 📷 for image · 🎤 for voice