Tech · No. 41 of 42 by public interest
Top AI models for coding, on Arena’s WebDev leaderboard
As of 28 Oct 2026, Claude Opus 5.5 (Max) is No. 1 in Arena’s WebDev leaderboard (1,810), ahead of GPT-6 Astra (Max) (1,773).
AI models ranked by Arena score on its WebDev leaderboard, where people ask two anonymous models to build the same web app and vote for the better one: the top 20.
- As of
- 28 Oct 2026
- Last changed
- 2 h ago
- Last checked
- 2 h ago
- Checked
- every 6 hours
Source: Arena, via Hugging Face · CC BY 4.0 · source data Shows the top 20 of the WebDev leaderboard, scores rounded.
The list
No earlier list to compare with yet
| No. | Move | Model | Arena score |
|---|---|---|---|
| 1 | Claude Opus 5.5 (Max)Anthropic | 1,810 | |
| 2 | GPT-6 Astra (Max)OpenAI | 1,773 | |
| 3 | Claude Sonnet 5.5 (xHigh)Anthropic | 1,770 | |
| 4 | Gemini 4 Carbon (High)Google | 1,768 | |
| 5 | GPT-6.1 Sol (Max)OpenAI | 1,751 | |
| 6 | Claude Fable 5.1 (Max)Anthropic | 1,746 | |
| 7 | Muse Wonder Full (Max)Meta | 1,725 | |
| 8 | Claude Sonnet 5.5 (High)Anthropic | 1,711 | |
| 9 | Kimi K3.1 Pro (Max)Moonshot AI | 1,695 | |
| 10 | Claude Opus 5 (Max)Anthropic | 1,691 | |
| 11 | Gemini 4 Argon (High)Google | 1,679 | |
| 12 | GPT-6 Sol (Max)OpenAI | 1,678 | |
| 13 | Qwen3.8-Max (0902)Alibaba | 1,676 | |
| 14 | GPT-6.1 Terra (Max)OpenAI | 1,673 | |
| 15 | Qwen3.8-MaxAlibaba | 1,672 | |
| 16 | DeepSeek-V4.1-Flash-20261017 (Max)DeepSeek | 1,665 | |
| 17 | Claude Opus 5 (High)Anthropic | 1,658 | |
| 18 | Kimi K3 (Max)Moonshot AI | 1,656 | |
| 19 | Muse Spark 1.3 (Max)Meta | 1,656 | |
| 20 | Muse Wonder (Max)Meta | 1,652 |
What changed
No changes recorded yet. The first change will appear here.