WebChoreArena - GitLab: leaderboard
Metric: Task success rate (%). Source: webchorearena.github.io. Saturation forecast: Rough model projection: around 2026. 6 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | BrowserGym + gemini-2.5-pro-preview-03-25 | 40.2 |
| 2 | AgentOccam + gemini-2.5-pro-preview-03-25 | 38.6 |
| 3 | BrowserGym + claude-3-7-sonnet-20250219 | 25.2 |
| 4 | AgentOccam + claude-3-7-sonnet-20250219 | 22.8 |
| 5 | AgentOccam + gpt-4o-2024-05-13 | 7.1 |
| 6 | BrowserGym + gpt-4o-2024-05-13 | 3.9 |
Interactive version: theaggregate.ai/benchmark?slug=webchorearena-gitlab · How It Works · Data refreshed daily, snapshot 2026-09-26.