Getting on Live Benchmarks

[BradKML]

There are two example benchmarks on what I am talking about, that Trae might want to check along side SWEBench and similar static benchmarks

• Ones that compare IDE tools https://liveswebench.ai/
• Ones that compare headless agents https://swe-bench-live.github.io/

[chao-peng]

Yeah, I am currently working on evaluation harness for Terminal Bench and SWE-bench Live.