WebTestBench - Functionality: leaderboard
Metric: Defect-detection F1 (%) on functionality test items, on 100 generated web applications, WebTester agents on Claude Code with Playwright MCP, predicted test items matched to the human gold checklist by a Qwen3.5-27B judge; higher is better. Source: arxiv.org. Saturation forecast: Around 2029. 10 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.1 | 30.9 | #131 |
| 2 | GPT-5.2 | 25.3 | #105 |
| 3 | Claude Sonnet 4.5 | 22.2 | #138 |
| 4 | MiMo-V2-Flash | 21.9 | #359 |
| 5 | Step 3.5 Flash | 20.1 | #262 |
| 6 | Claude Opus 4.5 | 18.8 | #79 |
| 7 | GLM-4.7 | 16.5 | #185 |
| 8 | Qwen3 Coder Next | 14.1 | #321 |
| 9 | MiniMax-M2.1 | 12.3 | #283 |
| 10 | GLM-5 | 11.9 | #137 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=webtestbench-functionality · How It Works · Data refreshed daily, snapshot 2026-10-11.