Pool
Pool — physically simulated billiards game.
What I asked each model — the Pool prompt
Every model on this page got this exact prompt inside the Agent Operating System: Pool — physically simulated billiards game.
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. 22 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, Grok, Inkling, Kimi K3, MiniMax M3, Hermes MoA, Opus 4.8, Claude Opus 5, Qwen 3.7, Claude Sonnet 5, Kimi K2.7 · Fast, Kimi K2.7 · No-Think, Kimi K2.7 · Quality, DeepSeek V4 Pro, DeepSeek V4 Flash, Kimi K2.7.
Why this task matters. Pool 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 Pool
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: Renders a clean 3D table with wood rails, six pockets, full 15-ball rack and cue ball, plus solid drag-to-aim physics and ball-ball collisions; slightly generic camera angle and the rack has spread post-autoplay-shot, but it's polished and shippable.
What I saw: Ultra v2 — billiards. Smoke-test MAYBE (0.0% diff) — cue needs click-drag aim that generic input didn't replicate.
What I saw: Visually polished 3D pool-hall arena with a giant table, pockets, minimap and combat HUD, but this is an arena shooter reskin ('cue-knight', enemies, HP/boost) — not a physically simulated billiards game as the brief demands, so it misses the core task entirely despite decent rendering.
What I saw: Canvas2D billiards on a green felt table with 16 balls, side + corner pockets, click-drag aim + power. Sound physics, on-brief.
What I saw: Visually polished 3D scene with working combat/enemies and particle FX, but it fundamentally misses the brief — this is a 'Rogue Pool Arena' shooter with phantoms and HP bars, not a physically simulated billiards game with proper cue mechanics and pocketed balls.
What I saw: Beautifully rendered table with realistic wood rail, felt gradient, numbered balls in a proper triangle rack, and clean HUD; physics/audio and pocketing logic are solid, though the presentation is more of an aesthetically strong standard billiards sim than a genre-redefining winner.
What I saw: Canvas2D billiards with 16 balls, pockets, click-drag aim + power. 16KB.
What I saw: Renders a clean 3D table with rack, pockets, and a decent title overlay, but the table sits small in the frame with heavy vignette wasting most of the screen, and the physics have flaws (cue-ball-only drag, no game rules/scoring). The rack looks slightly offset and the balls are plain-colored with no numbers/stripes, keeping it generic rather than a task winner.
What I saw: Renders a polished 3D pool table with proper rack, numbered balls, cue stick, aim line, pockets, HUD and physics/audio scaffolding — clearly shippable and on-brief. Weak points: the floating lamp shade looks unlit/misplaced and the rack triangle is slightly loose, keeping it just below the top tier.
The winner on Pool
Hermes MoA took gold on this task. hermes.
What I saw: Polished Canvas2D billiards with full 16-ball physics, substepped collision resolution, pocket-suction zones, scratch respotting, auto-break, particle effects and a clean drag-power cue with predictive line — clearly edges Fusion/Grok (8.0) on physics fidelity and presentation, and far ahead of SOLO Opus 4.8 (5.5) which had no input response. Just shy of a flawless top tier (simple normal-impulse collisions, no friend/foe ball rules), but easily the strongest on this task.
See Hermes MoA's full model card: /models/moa. Direct head-to-head against the runner-up: Hermes MoA vs GPT-5.6 Sol.
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
▶ LIVEHow I scored Pool — 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 Pool · 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.