Parallel WebBench - Binary Accuracy (WebExplorer-Style Scaffold): leaderboard
Metric: Binary accuracy (%; GPT-4.1-mini judge, whole structured answer correct) on the 150 held-out tasks, 16k input and 16 interaction rounds, WebExplorer-Style Scaffold. Source: arxiv.org. Saturation forecast: Around April 2027. 6 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-4.1 Mini | 28 |
| 2 | Qwen 2.5 7B Instruct | 8 |
Interactive version: theaggregate.ai/benchmark?slug=parallel-webbench-binary-accuracy-webexplorer-style-scaffold · How It Works · Data refreshed daily, snapshot 2026-09-26.