Crypt
Crypt — torch-lit dungeon crawler.
What I asked each model — the Crypt prompt
Every model on this page got this exact prompt inside the Agent Operating System: Crypt — torch-lit dungeon crawler.
Single HTML file out. No iteration. No examples in the system prompt. Whatever each model produced on the first run is what's on this page. 24 frontier models have attempted it so far: Claude Fable 5, Fugu Ultra, Fugu Ultra 1.1, Fusion, Gemini 3.6 Flash, GLM-5.2, GPT-5.6 Sol, Inkling, Kimi K3, LongCat-2.0, MiniMax M3, Hermes MoA, Opus 4.8, Claude Opus 5, Qwen 3.8, Qwen 3.7, Claude Sonnet 5, DeepSeek V4 Pro, DeepSeek V4 Flash, Grok, Kimi K2.7, Kimi K2.7 · Fast, Kimi K2.7 · No-Think, Kimi K2.7 · Quality.
Why this task matters. Crypt is a textbook test of game-class capability — the kind of build that exposes whether a model is doing pattern-matching or actual reasoning. A model that ships this in one shot is usually safe to wire into your agent loop for harder tasks of the same shape.
How each model handled Crypt
Ranked by my 0–10 score from the source comparison guides on agentos.guide. Click any to play the actual one-shot HTML the model produced.
What I saw: AAA rebuild via the threejs-game-director skill transforms it into a real torch-lit crawler: warm-lit cracked-stone walls, a stone sarcophagus, a flaming torch held in first-person, scattered rubble and soul-gem pickups, drifting dust, and a cohesive HUD (vitality, objective + distance, compass, gem counter). WASD/mouse-look verified. Needed a one-line patch to a glitched emissive token and the interior reads a touch boxy, keeping it just under the very top.
What I saw: Ultra v2 (gap-fill) — first-person Nordic dungeon. Smoke-test MAYBE (0.4% diff) — pointer-lock FPS; flagged for manual verification.
What I saw: Strong atmospheric torch-lit crypt with a visible skeleton enemy, working HUD (health/torch/kills/depth), coffins and props — clearly a combat crawler not an empty walk sim. Lighting feels a bit washed-out/bright for a 'torch-lit dungeon' and the mood lacks the deep shadowy dread of the top entry, keeping it shippable but not quite a task winner.
What I saw: First-person Nordic dungeon on three.js with PointerLockControls + WebGL. Torch-lit corridors, held torch, skeletons to strike, health + gold HUD. The crypt Julian wanted.
What I saw: Polished HUD with vitality/stamina/torch bars, viewmodel sword and torch, and clean brick dungeon rendering, but the screenshot shows no visible enemies or combat and the lighting reads flat/brown rather than moody torch-lit — a competent but empty walking sim so far.
What I saw: 29KB · plays clean · three, webgl, pointer-lock
What I saw: Renders a clean, atmospheric 3D crypt with textured stone walls, pillars, archways, a floating rune and full HUD (torch bar, rune counter, controls) — a solid, shippable first-person crawler. Weakness: lighting reads more purple-lavender than 'torch-lit,' the ambient wash flattens the dungeon mood and undercuts the flickering-flame promise, keeping it just short of the field's best.
What I saw: The 3D scene fails to render — the canvas is entirely black with only the UI overlay (title, hint, controls) visible, because the code calls `wallGeo(w,h,d)` as a function when `wallGeo` is a BoxGeometry instance, throwing an error that halts scene setup. Nice atmospheric typography and control styling, but the dungeon crawler itself is broken.
What I saw: Renders a clean, atmospheric brick corridor with polished title typography, a working minimap, and solid HUD framing, but the scene reads cool-blue and under-lit rather than warmly torch-lit — no visible flames or orange glow undercut the core brief, keeping it strong-but-generic rather than a task winner.
What I saw: One-shot 9KB torch-lit stone dungeon corridor — pillars, barrels, a chest, 6+ flickering torch PointLights, fog. Real WASD+mouse controls. verify-move: walks+looks, 0 errors. Lit + atmospheric (a touch over-bright orange).
The winner on Crypt
Fusion took gold on this task. winner · dungeon crawler.
What I saw: First-person Nordic dungeon on three.js with PointerLockControls + WebGL. Torch-lit corridors, held torch, skeletons to strike, health + gold HUD. The crypt Julian wanted.
See Fusion's full model card: /models/fusion. Direct head-to-head against the runner-up: Fusion vs Claude Fable 5.
Every attempt — live, playable
Side by side. Click any tile to run that model's actual one-shot HTML in a new tab.
▶ LIVE
▶ LIVE
▶ LIVE
▶ LIVE
▶ LIVE
▶ LIVE
▶ LIVE
▶ LIVE
▶ LIVE
▶ LIVE
▶ LIVE
▶ LIVE
▶ LIVE
▶ LIVE
▶ LIVE
▶ LIVE
▶ LIVE
▶ LIVE
▶ LIVE
▶ LIVE
▶ LIVE
▶ LIVE
▶ LIVE
▶ LIVEHow I scored Crypt — methodology
Three axes, 0–10 each, averaged. Runs: drop the .html in a browser; if it opens to a broken page, it scores zero. Hits the brief: did the model ship the thing the prompt asked for, or a different thing it found easier. Looks good: visual polish, motion, interactivity — where most of the gap between gold and silver lives.
My scores trace back to the source comparison guides on agentos.guide. See the full methodology page for data provenance, including which source guide each cell's score came from.
Related
More game benchmarks: all tasks in the Game category · See the best AI model for Crypt · Back to the leaderboard
Run this stack yourself.
Every demo on this bench was built inside the Agent Operating System — one prompt, one shot, single HTML file out. The Agent OS, the prompts, the templates, the weekly walkthroughs and 4,000+ founders shipping with it every day all live inside the AI Profit Boardroom.