LMArena WebDev Arena: leaderboard
Human voters submit a prompt for a web app, two anonymous models build it, and both builds run in isolated sandboxes the voter can click through before picking one. This board is the overall WebDev rating across all categories, mirrored from benchmarklist.com, which imports LMArena's published leaderboard data. Votes are fitted with a Bradley-Terry model and higher is better. Source: arena.ai/blog/webdev-arena.
Metric: Arena rating (imported). Source: huggingface.co. Source: benchmarklist.com. Saturation forecast: Around December 2027. 35 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Fable 5.1 (Max) | 1765.37 | #9 (Claude Fable 5.1) |
| 2 | Qwen 3.8 Max (0902) | 1687.74 | #29 |
| 3 | Claude Opus 5 (Max) | 1687.23 | #12 (Claude Opus 5) |
| 4 | Kimi K3 (Max) | 1673.82 | #30 (Kimi K3) |
| 5 | Qwen 3.8 Max | 1669.22 | #35 |
| 6 | Claude Opus 5 (High) | 1661.23 | #12 (Claude Opus 5) |
| 7 | Grok 4.6 (High) | 1629.06 | #26 (Grok 4.6) |
| 8 | Claude Fable 5 | 1628.42 | #16 |
| 9 | Hy4 preview | 1625.78 | #38 |
| 10 | Qwen3.8-Flash-Next | 1622.28 | #72 |
| 11 | GPT-5.6 Sol (xHigh) | 1616.41 | #19 (GPT-5.6 Sol) |
| 12 | GLM-5.3 (Max) | 1609.17 | #39 (GLM-5.3) |
| 13 | GLM-5.3 Flash | 1604.43 | #65 |
| 14 | Qwen 3.8 27B | 1597.46 | #105 |
| 15 | Gemini 3.7 Flash (High) | 1586.62 | #23 (Gemini 3.7 Flash) |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=lmarena-webdev-arena · How It Works · Data refreshed daily, snapshot 2026-10-20.