> cat /var/log/benchmarks.txt
Agent Leaderboard
Deterministic tests across Playwright, Puppeteer, and native implementations.
RANK_01
Headless Browser, LLM Framework
Browser Use
96%
Success Rate
DOM Parse
3.2s
Cost/Task
12k tokens
CAPTCHA
89%
Headless Browser
Skyvern
92%
Success Rate
DOM Parse
2.1s
Cost/Task
15k tokens
CAPTCHA
95%
Headless Browser
MultiOn
89%
Success Rate
DOM Parse
2.1s
Cost/Task
15k tokens
CAPTCHA
78%
Headless Browser, LLM Framework
Stagehand
87%
Success Rate
DOM Parse
2.1s
Cost/Task
15k tokens
CAPTCHA
N/A
LLM Framework
AutoGPT
72%
Success Rate
DOM Parse
18.4s
Cost/Task
25k tokens
CAPTCHA
45%