Blind arena pits 6 AI models on the same voxel space station; build times span 5 min to 2h20
Miserable_Tiger_5742 · reddit · 2026-10-17
A developer built a blind arena page where six models get the exact same prompt, run once with no edits, and users vote before seeing who made what. This round: an explorable voxel space station on desktop and phone, built by Opus 5.5, Sonnet 5.5, Fable 5.1, Muse Wonder 1.0, GPT-6.1-Sol and GPT-6-Astra at max effort. Build times ranged from under 5 minutes to 2h20m; five models used the same 3D library while one wrote its own engine from scratch, and one filled its station with voxel villagers. The arena itself was built by Claude (Cloudflare Pages Functions + KV). Runs happen in Locust, an MIT-licensed open-source desktop app for running Claude Code, Codex and others side by side. Try it free at locust.lol/arena/space.
Related event: AI Models Blind-Tested on Building a Voxel Space Station(2 posts)→
More from coding & agent
- Cognition open-sources Devin agent plugin to patch the day's CVE wave — Scobleizer · 2026-10-23
- Fine-tuning models solo with a Grok bot and rented GPUs — Rasmic · 2026-10-23
- One prompt found pre-auth PHP RCEs: agent-guided vuln hunting timeline — chrisrohlf · 2026-10-23
- AutoRouter launches beta: auto-routes each coding task across Claude, GPT and other models to cut token spend — floguo · 2026-10-23
- How Speak uses GPT-Live and a Decisions API for ultra-responsive live AI language tutoring — pbbakkum · 2026-10-23
- Google Live API adds decisions model triggers, faster than function calls for voice agents — pbbakkum · 2026-10-23