τ^τ-Bench: An Environment for End-To-End, Realistic Agent Construction
New τ^τ-bench tasks coding agents with building deployable customer-service agents; best config, Claude Opus 5, passes only 23.9% of simulations.
1 stories, newest first · one card per story
New τ^τ-bench tasks coding agents with building deployable customer-service agents; best config, Claude Opus 5, passes only 23.9% of simulations.
Today 1K readers · 1.5K views · 1.1K reads · all time 159K views