Get the Agent OS + join 4,000+ founders inside the AI Profit Boardroom → Join AIPB ($59/mo)
Sim

Fractal

Fractal — interactive fractal explorer (mandelbrot or julia).

CategorySim
Models tested23
Scored18/23
Avg score8.08/10
WinnerKimi K2.7

What I asked each model — the Fractal prompt

Every model on this page got this exact prompt inside the Agent Operating System: Fractal — interactive fractal explorer (mandelbrot or julia).

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. 23 frontier models have attempted it so far: Claude Fable 5, Fugu Ultra, Fugu Mini, Fusion, Gemini 3.6 Flash, GLM-5.2, GPT-5.6 Sol, Grok, Inkling, Kimi K2.7, Kimi K3, 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, Kimi K2.7 · Fast, Kimi K2.7 · No-Think, Kimi K2.7 · Quality.

Why this task matters. Fractal is a textbook test of sim-class capability — the kind of build that exposes whether a model is doing pattern-matching or actual reasoning. Shipping this cleanly is the floor for what I expect from a frontier model — every model on the leaderboard should at least attempt it.

How each model handled Fractal

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.

Claude Fable 5 Anthropic
• 8.4/10 · Crisp GPU Mandelbrot

What I saw: Renders a clean, high-iteration Mandelbrot with sharp filament detail and a rich red/blue gradient, plus full pan/zoom/pinch, Julia toggle, morphing constant, and autoplay zoom. Very strong and shippable; slightly generic palette and the deep-zoom float precision limit keep it just under the top tier.

▶ Play Claude Fable 5's attempt →
Fugu Ultra Sakana AI
• 8.0/10

What I saw: Ultra v2 — Mandelbrot zoom. Smoke-test PASS.

▶ Play Fugu Ultra's attempt →
Fugu Mini Sakana AI
• 6.5/10

What I saw: Mandelbrot fractal. Smoke-test MAYBE-STATIC — visible but click-zoom may not be wired.

▶ Play Fugu Mini's attempt →
Fusion OpenRouter
• 8.5/10

What I saw: WebGL Mandelbrot shader with click-to-zoom and hold-to-zoom-continuously. HUD shows current center re/im coordinates and zoom level live. Mode selector for color schemes, kbd-styled key hints, flash transitions. Beats most fractal attempts on UI feedback.

▶ Play Fusion's attempt →
• 8.6/10 · GLSL glow Mandelbrot

What I saw: Gorgeous WebGL-shader Mandelbrot with striking neon glow rays, orbit-trap detail and a full control suite (type toggle, palettes, iteration/depth sliders, auto-morph, HUD) that reads as premium and interactive. Only minor concern is the 9 FPS reading suggesting heavy load, but the visual polish and completeness rival the field's best.

▶ Play Gemini 3.6 Flash's attempt →
GLM-5.2 Zhipu / Z.ai
• 8.0/10

What I saw: All three are genuinely good. Kimi's is the jaw-dropper — a deep rainbow plunge into a seahorse spiral, dense with self-similar detail. Opus zooms smoothly into the seahorse valley with a tasteful cycling palette. GLM frames the whole iconic set in a fire palette with a live coordinate HUD, then descends. Kimi takes this one on raw spectacle.

▶ Play GLM-5.2's attempt →
GPT-5.6 Sol OpenAI
• 8.6/10 · polished GPU Mandelbrot

What I saw: Crisp WebGL-rendered Mandelbrot with a striking crimson-to-cyan palette, clean glass HUD showing live zoom/coordinates, and full pan/zoom/animate controls plus a canvas fallback; the classic centered set is well-composed and edge detail is sharp, making it a task-topping build. Minor nit: palette is vivid but somewhat conventional and could feel busy at deep zoom.

▶ Play GPT-5.6 Sol's attempt →
Grok xAI
• 7.5/10

What I saw: A clean, vivid Mandelbrot — the classic shape in smooth rainbow colour, with an auto-zoom and scroll-to-zoom. Renders reliably now.

▶ Play Grok's attempt →
Inkling Thinking Machines
• 7.8/10

What I saw: Clean GPU-shader Mandelbrot renders crisply with smooth iteration coloring, glow, and polished title/hint overlay plus proper pan+zoom with lerp smoothing. Loses ground on being pan-only (no zoom-to-cursor), a slightly muddy blue-orange palette, and lacking Julia/coordinate readouts that top entries offered.

▶ Play Inkling's attempt →
Kimi K2.7 Moonshot AI
🥇 9.0/10 · winner · pure wow

What I saw: All three are genuinely good. Kimi's is the jaw-dropper — a deep rainbow plunge into a seahorse spiral, dense with self-similar detail. Opus zooms smoothly into the seahorse valley with a tasteful cycling palette. GLM frames the whole iconic set in a fire palette with a live coordinate HUD, then descends. Kimi takes this one on raw spectacle.

▶ Play Kimi K2.7's attempt →

The winner on Fractal

Kimi K2.7 took gold on this task. winner · pure wow.

What I saw: All three are genuinely good. Kimi's is the jaw-dropper — a deep rainbow plunge into a seahorse spiral, dense with self-similar detail. Opus zooms smoothly into the seahorse valley with a tasteful cycling palette. GLM frames the whole iconic set in a fire palette with a live coordinate HUD, then descends. Kimi takes this one on raw spectacle.

See Kimi K2.7's full model card: /models/kimi. Direct head-to-head against the runner-up: Kimi K2.7 vs Hermes MoA.

Every attempt — live, playable

Side by side. Click any tile to run that model's actual one-shot HTML in a new tab.

How I scored Fractal — 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.

The same stack Julian uses

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.

4,000+founders
258documented wins
38countries
$59/momonthly