Interesting benchmark, but I'd be careful with the headline number. Each target looks like a single url, so 62.25% is an average of four data points, and most of the gap comes from tripadvisor and amazon. It's also default against default, since browserbase ran fetch with no js and proxies off, so it's hard to say what it can do when configured properly
The cost numbers seem to price a workload that wasn't benchmarked, and the 40% failure rate doesn't quite match the article's own cloudflare target, where browserbase got 100%. It also isn't clear what counts as success, validated content or just a 200
Interesting benchmark, but I'd be careful with the headline number. Each target looks like a single url, so 62.25% is an average of four data points, and most of the gap comes from tripadvisor and amazon. It's also default against default, since browserbase ran fetch with no js and proxies off, so it's hard to say what it can do when configured properly
The cost numbers seem to price a workload that wasn't benchmarked, and the 40% failure rate doesn't quite match the article's own cloudflare target, where browserbase got 100%. It also isn't clear what counts as success, validated content or just a 200