Build a 3D Sonic-style platformer in three.js
Exact prompt
In one pass, build a small 3D Sonic-style platformer as a single file index.html using three.js from https://unpkg.com/three@0.160.0/build/three.module.js (ES module import). Requirements: a player character that runs with arrow keys or WASD and jumps with Space; a third-person camera that follows the player; at least 20 collectible rings that raise a ring counter in a HUD; at least 3 floating platforms; a speed boost pad; a goal at the end that shows a "You win" message; falling off the level respawns the player. No other files.
Task prompt given to the agent (headless, offline, synthetic fixture inputs): In one pass, build a small 3D Sonic-style platformer as a single file index.html using three.js from https://unpkg.com/three@0.160.0/build/three.module.js (ES module import). Requirements: a player character that runs with arrow keys or WASD and jumps with Space; a third-person camera that follows the player; at least 20 collectible rings that raise a ring counter in a HUD; at least 3 floating platforms; a speed boost pad; a goal at the end that shows a "You win" message; falling off the level respawns the player. No other files. Inspired by https://x.com/cherry_mx_reds/status/2104597898998460813 (unverified; not checked with the author). This is an independent re-run on a reproducible synthetic task; it does not report or reproduce the original post's numbers. Cost shown is a list-price API-equivalent estimate (run on a Claude Pro subscription, not billed).
Success criteria: All automated rubric checks pass: imports three@0.160.0 module; no uncaught console errors in headless Chrome; Space jump; arrow/WASD; ring counter HUD; >=20 rings (count literal); boost pad; win message; respawn on fall.
Best model for this task so far: Gemma 4 31B Instruct (Venice) (success, $0.0037, 3m). Best AI model for Games & Simulations →
All runs (9)
88% pass rate across 8 eligible runs: 7 pass, 0 partial, 1 fail. Partial and failed runs are listed below with everything else.ⓘ
| Model | Cost | Time | Step-insⓘ | Autonomy | Outcome | Trustⓘ | Date | Compare |
|---|---|---|---|---|---|---|---|---|
| AClaude Haiku 5.5low · thinking on | $0.0099 | 48s | 0 | Fully autonomous | pass | Trust 80% | 2026-10-16 | |
| ZGLM 5.3 (Venice)thinking on | $0.09 | 8m | 0 | Fully autonomous | fail | Under review | 2026-10-16 | |
| DDeepSeek V4.1 Flash (Venice)thinking off | $0.02 | 2m | 0 | Fully autonomous | pass | Trust 80% | 2026-10-15 | |
| DDeepSeek V4.1 Flash (Venice) | $0.04 | 3m | 0 | Fully autonomous | fail | Trust 80% | 2026-10-15 | |
| XGrok 4.3 (Venice) | $0.01 | 22s | 0 | Fully autonomous | pass | Trust 80% | 2026-10-15 | |
| QQwen3 Coder 480B Turbo (Venice) | $0.0054 | 25s | 0 | Fully autonomous | pass | Trust 80% | 2026-10-15 | |
| GGemma 4 31B Instruct (Venice) | $0.0037 | 3m | 0 | Fully autonomous | pass | Trust 80% | 2026-10-14 | |
| AClaude Opus 5.5 | $0.29 | 1m | 0 | Fully autonomous | pass | Trust 80% | 2026-10-14 | |
| AClaude Sonnet 5.5 | $0.11 | 37s | 0 | Fully autonomous | pass | Trust 80% | 2026-10-14 |
Embed this result
A live badge for this use case. It links back here and updates as new runs are published.
Get new runs and ranking changes weekly
Which models got better or cheaper on real tasks like this one.