Arena Image-to-WebDev
Code
Arena Image-to-WebDev ranks models that turn visual references into working web experiences. Anonymous generated results are compared through blind pairwise user votes.
Top models (higher is better)
| Model | Score |
|---|---|
| Claude Opus 5 | 1669 |
| Claude Fable 5 | 1626 |
| Opus 4.7 | 1582 |
| Grok 4.5 | 1578 |
| Kimi K3 | 1571 |
| Muse Spark 1.1 | 1550 |
| Opus 4.6 | 1543 |
| Gemini 3.6 Flash | 1543 |
| Sonnet 5 | 1540 |
| Sonnet 4.6 | 1529 |
| Kimi K2.6 | 1521 |
| Kimi K2.7 Code | 1513 |
| Gemini 3.5 Flash | 1493 |
| Gemini 3.1 Pro Preview | 1481 |
| MiniMax M3 | 1479 |
| Qwen3.6 Plus (2026-04-02) | 1471 |
| Gemini 3 Flash Preview | 1456 |
| Gemini 3 Pro Preview | 1447 |
| GPT-5.4 | 1447 |
| Kimi K2.5 | 1441 |
| Gemini 3.5 Flash-Lite | 1437 |
| GLM 5V Turbo | 1424 |
| GPT-5.1 | 1420 |
| Grok 4.3 | 1377 |
| Gemini 3.1 Flash-Lite Preview | 1337 |
| Gemini 2.5 Pro | 1283 |