Announcing AA-3D: Artificial Analysis benchmarks AI models building interactive 3D scenes
ArtificialAnlys · x · 2026-10-13
Artificial Analysis launched AA-3D, a benchmark measuring how well AI models build interactive 3D scenes people want to explore. Models act as agents in an offline sandbox (Blender, Rapier physics, headless browser, CC0 assets), turning a written brief into a three.js page, reviewing their own screenshots and revising; a human panel blind-compares pairs and votes become Elo via Bradley-Terry. Claude Sonnet 5.5 leads at 1194, statistically tied with three other Anthropic/OpenAI models; Meta's Muse Spark 1.3 is seventh at 1035; best open-weights models Kimi K3 (982) and MiMo-V2.6-Pro (977) rank 11th–12th, 200+ points behind. A live leaderboard and a public blind-voting page accompany the launch.
Related event: Artificial Analysis Launches AA-3D Benchmark for Interactive 3D Models(4 posts)→
More from Models
- Anthropic says it cut distiller data harvesting by 90-99%, turning away trillions of tokens — abhiadesai · 2026-10-13
- Ex-exMVP: post-training should add a facet for how well models explain things — MParakhin · 2026-10-13
- GLM-5.3 abliterated for a month, no takeover: dev rails against AI doom narratives — mariofilhoml · 2026-10-13
- Claude auto-files a bug report about itself: TLDRs still too long after user asked for shorter — andrey_kurenkov · 2026-10-13
- BridgeBench one-prompt test shows Muse Wonder is fast and cheap but worst-looking — PVORY · 2026-10-13
- Meta skipped the 'too dangerous to release' marketing for Muse Wonder — and could rival Google — haider1 · 2026-10-13