Four Models Play Doom

Real-time comparison of AI models playing Doom via vision-to-action text output

Description

A vision-to-action demo comparing the performance of four AI models—Jev 1.13, Laya, ModernCE, and Qwen3.5—in the Doom 'Defend the Center' scenario. The system operates at a 35Hz world rate with a 5Hz target, demonstrating that small local models can achieve performance parity with hosted flagship APIs using pure text output controls.

Descriptions, tags, and model credits may be AI-generated or inferred from public sources and can be incomplete or wrong. Learn more.

Details

Why this is interesting: Demonstrates impressive capability by showing small local models competing with flagship AIs in a complex game…
Attributed to: @himanshu231204
Built with: Redwood AI Jev Laya ModernCE-base Qwen3.5 4B
Tags: #doom #ai-agents #benchmarking #vision-to-action

Discovered: X
First posted: X · Sep 25, 2026
Listed: Added 16d ago
Build: Strong
Models: Strong