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

Racing

3D Racer — third-person racing game with a track and obstacles.

CategoryGame
Models tested24
Scored18/24
Avg score7.80/10
WinnerFusion

What I asked each model — the Racing prompt

Every model on this page got this exact prompt inside the Agent Operating System: 3D Racer — third-person racing game with a track and obstacles.

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, Fugu Mini, 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 3.8, 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. Racing 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 Racing

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
• 8.3/10

What I saw: Strong, clean third-person 3D racer with proper perspective road, curbs, trees, colorful obstacles, follow-cam, HUD, lives and touch controls all rendering crisply. Solid and shippable, but visuals are somewhat generic/flat and it lacks a distinctive polish edge (e.g., turns, minimap, real track curves) to top the field's best.

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

What I saw: Ultra v2 — 3D arcade racer with drift. Smoke-test PASS.

▶ Play Fugu Ultra's attempt →
Fugu Ultra 1.1 Sakana AI
• 8.6/10 · Neon combat racer

What I saw: Gorgeous polished third-person racer with a clear track, guardrails, trees/rocks/buildings, obstacle cones, a slick craft with ground shadow, and combat layered in (crosshair, KILLS 0/12, visible enemies/targets ahead) plus rich HUD with minimap — strong shippable build; only minor knock is kills still at 00 in this frame, but the enemy/combat system is clearly present and working.

▶ Play Fugu Ultra 1.1's attempt →
Fugu Mini Sakana AI
• 8.0/10

What I saw: 3D arcade racer with banking turns + drift. Smoke-test PASS.

▶ Play Fugu Mini's attempt →
Fusion OpenRouter
🥇 9.0/10 · winner · arcade racer

What I saw: RETRY @ 24K tokens — now complete: 32KB three.js with rAF + 4 input handlers + closed tags. Banking turns, drift, obstacles, lap timer. The original truncated attempt has been replaced.

▶ Play Fusion's attempt →
• 4.2/10

What I saw: Polished neon HUD with speed, laps, minimap, health/nitro and combat labels, but the rendered scene shows only floating dark slabs with no visible track, car, or enemies — the core third-person racer and combat aren't apparent on screen. Strong UI framing wasted on an incoherent, near-broken 3D view.

▶ Play Gemini 3.6 Flash's attempt →
GLM-5.2 Zhipu / Z.ai
• 7.5/10

What I saw: 34KB · plays clean · plain

▶ Play GLM-5.2's attempt →
GPT-5.6 Sol OpenAI
• 8.4/10

What I saw: Strong, clean third-person 3D racer with a well-modeled player car, receding dashed-line track, guardrails, low-poly trees, an AI car and cones as obstacles, plus a polished HUD (distance/speed/boost). Cohesive lighting and shadows make it shippable, but it reads slightly generic/soft next to the field's best and stops just short of a clear task winner.

▶ Play GPT-5.6 Sol's attempt →
Grok xAI
• 8.5/10

What I saw: 3D arcade racer, third-person, banking turns, drift mechanic, obstacles, lap timer. 29KB on second retry.

▶ Play Grok's attempt →
Inkling Thinking Machines
• 5.2/10

What I saw: Strong neon title/HUD polish and the car mesh renders cleanly, but the track is essentially invisible in the dark scene — it reads as floating boxes in space rather than a road, undermining the core 'track and obstacles' brief. Functional third-person camera and physics in code, but the visual presentation looks broken/unclear compared to the field's best.

▶ Play Inkling's attempt →

The winner on Racing

Fusion took gold on this task. winner · arcade racer.

What I saw: RETRY @ 24K tokens — now complete: 32KB three.js with rAF + 4 input handlers + closed tags. Banking turns, drift, obstacles, lap timer. The original truncated attempt has been replaced.

See Fusion's full model card: /models/fusion.

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 Racing — 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