Arena Code WebDev — Reference-Based Design
Code
Arena Code WebDev rating for reference-based design tasks. Anonymous generated experiences are compared through blind pairwise user votes.
Top models (higher is better)
| Model | Score |
|---|---|
| Claude Opus 5.5 | 1837 |
| Claude Sonnet 5.5 | 1824 |
| GPT-6 Astra | 1807 |
| Claude Fable 5.1 | 1800 |
| GPT-6.1 Sol | 1779 |
| Muse Wonder Full | 1725 |
| GPT-6 Sol | 1720 |
| Claude Opus 5 | 1708 |
| Gemini 4 Argon | 1703 |
| Qwen3.8 Max (0902) | 1688 |
| Qwen3.8-Max | 1685 |
| Kimi K3.1 Pro | 1684 |
| Kimi K3 | 1679 |
| Muse Spark 1.3 | 1672 |
| DeepSeek V4.1 Flash 1017 | 1669 |
| Muse Wonder | 1660 |
| Claude Fable 5 | 1655 |
| GPT-5.6 Sol | 1653 |
| Grok 4.7 | 1651 |
| Hy4 Preview | 1650 |
| MiMo-V2.6-Pro | 1644 |
| MiMo-V2.6-Flash | 1644 |
| GLM-5.3 | 1640 |
| Grok 4.6 | 1635 |
| MiniMax M3.1 Flash | 1635 |
| Qwen3.8-Flash-Next | 1631 |
| GLM-5.3 Flash | 1615 |
| DeepSeek V4.1 Flash | 1614 |
| Gemini 3.7 Flash | 1608 |
| GLM-5.2 | 1608 |
| Claude Haiku 5.5 | 1606 |
| Qwen3.8 27B | 1597 |
| Gemini 3.8 Flash | 1596 |
| GPT-6 Luna | 1587 |
| DeepSeek V4 Pro 0813 | 1583 |
| Opus 4.7 | 1580 |
| DeepSeek-V4-Flash | 1577 |
| Opus 4.8 | 1577 |
| Grok 4.5 | 1575 |
| Opus 4.6 | 1561 |