WebDev Arena (Fullstack): leaderboard
Voters submit a prompt for a web app, two anonymous models build it, and both builds run in sandboxes the voter can click through before choosing one. This board is the fullstack slice, where the prompt calls for backend work and models can generate a PostgreSQL database layer with user authentication, using bash and web search tools. Votes are fitted with a Bradley-Terry model and higher is better.
Metric: Arena Score. Source: arena.ai. 78 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-6 Astra (Max) | 1756 | #6 (GPT-6 Astra) |
| 2 | Claude Opus 5.5 (Max) | 1741 | #3 (Claude Opus 5.5) |
| 3 | GPT-6 Sol (Max) | 1705 | #14 (GPT-6 Sol) |
| 4 | Claude Sonnet 5.5 (xHigh) | 1700 | #11 (Claude Sonnet 5.5) |
| 5 | Qwen 3.8 Max | 1689 | #35 |
| 6 | Gemini 4 Argon (High) | 1686 | #8 (Gemini 4 Argon) |
| 7 | Claude Fable 5.1 (Max) | 1679 | #9 (Claude Fable 5.1) |
| 8 | Qwen 3.8 Max (0902) | 1673 | #29 |
| 9 | Claude Opus 5 (Max) | 1666 | #12 (Claude Opus 5) |
| 10 | Claude Sonnet 5.5 (High) | 1659 | #11 (Claude Sonnet 5.5) |
| 11 | Kimi K3 (Max) | 1657 | #30 (Kimi K3) |
| 12 | Muse Spark 1.3 (Max) | 1653 | #24 (Muse Spark 1.3) |
| 13 | Claude Opus 5 (High) | 1635 | #12 (Claude Opus 5) |
| 14 | Muse Spark 1.3 (xHigh) | 1635 | #24 (Muse Spark 1.3) |
| 15 | GPT-5.6 Sol (xHigh) | 1634 | #19 (GPT-5.6 Sol) |
Interactive version: theaggregate.ai/benchmark?slug=webdev-arena-fullstack · How It Works · Data refreshed daily, snapshot 2026-10-20.