Terminal-Bench
Terminal-Bench 4.1 tightens grading for AI terminal agents
Terminal-Bench 4.1 is now available with stronger verifier isolation, reproducibility improvements and a 63-task CPU-only dataset. The release keeps the benchmark’s 66…
Replit lets its agent's model pick subagents, effort and delegation itself
Replit argues that model routers are limited by design, and says its Replit Agent instead lets the main model decide which subagents to use, how large they are and how…