StressWeb - Action Failure: leaderboard
StressWeb (2026) is a diagnostic stress test for web agents: generated websites with clean reference workflows are re-run under controlled perturbations, and deterministic scripts check task checkpoints on the recorded page states. This board is the checkpoint pass rate in percent when clicks and text submissions randomly fail; higher is better.
Metric: Checkpoint pass rate (%): share of predefined task checkpoints (milestones such as reaching a target page or completing required fields) that the agent's run passes, under execution perturbation in which clicks and text submissions randomly fail (probability 0.35), on 149 tasks over 10 generated websites, a multimodal agent driving a Playwright controller from screenshots, the page DOM and its action history (100-step budget), checkpoints verified by task-specific scripts on the recorded DOM states; higher is better. Source: arxiv.org. Saturation forecast: Around April 2027. 8 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | O4 Mini | 54 | #368 |
| 2 | Gemini 2.5 Pro | 51.3 | #276 |
| 3 | Claude Opus 4.5 | 47.7 | #160 |
| 4 | Claude Sonnet 4.5 | 47.5 | #261 |
| 5 | GPT-5.2 | 45.5 | #185 |
| 6 | Qwen 3 VL 235B A22B (Thinking) | 43.4 | #459 |
| 7 | GLM-4.6V | 40.1 | #710 |
| 8 | Claude Haiku 4.5 | 39.7 | #583 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=stressweb-action-failure · How It Works · Data refreshed daily, snapshot 2026-11-21.