Benchmarks · creative
threejseval
Coding models build a Three.js scene from the same prompt; humans vote blind on pairs of finished scenes. Value is Bradley-Terry Elo (base 1500), the mean across prompts.
30 models · 38 results · updated 2026-11-04 · threejseval.com ↗
Leaderboard
Opus 5.5 high
More from this benchmark
3 published measuresScore vs any published measure3 measures: sub-scores, splits, counts
16 more models under All models in settings.
All results
38 rows| # | Model | Effort | threejseval |
|---|---|---|---|
| 1 | Opus 5.5 | high | 2067 |
| 2 | GPT-6.1 Sol | high | 1973 |
| 3 | Opus 5.5 | medium | 1889 |
| 4 | GPT-6 Astra | medium | 1888 |
| 5 | Fable 5.1 | high | 1864 |
| 6 | Sonnet 5.5 | high | 1833 |
| 7 | DeepSeek V4.1 Flash 1017 | max | 1719 |
| 8 | Opus 5 | medium | 1713 |
| 9 | Gemini 4 Argon | high | 1707 |
| 10 | DeepSeek V4.1 Flash | max | 1634 |
| 11 | GPT-6 Sol | high | 1626 |
| 12 | Grok 4.7 | xhigh | 1622 |
| 13 | Grok 4.7 | high | 1598 |
| 14 | MiniMax M3.1 | max | 1591 |
| 15 | Grok 4.7 | medium | 1574 |
| 16 | Gemini 3.8 Flash | high | 1568 |
| 17 | Fable 5 | medium | 1517 |
| 18 | Kimi K3 | high | 1514 |
| 19 | Kimi K3.1 | high | 1511 |
| 20 | Gemini 3.7 Flash | high | 1469 |
| 21 | Kimi K3 | low | 1460 |
| 22 | Grok 4.6 | high | 1457 |
| 23 | Muse Spark 1.3 | max | 1438 |
| 24 | GPT-5.6 Terra | max | 1433 |
| 25 | Grok 4.7 | low | 1424 |
| 26 | Haiku 5.5 | high | 1412 |
| 27 | GPT-5.6 Sol | medium | 1371 |
| 28 | GPT-6 Luna | high | 1367 |
| 29 | Gemini 3.6 Flash | high | 1343 |
| 30 | Grok 4.6 | medium | 1286 |