Posts tagged with "Benchmarks"

τ-bench tests whether a model can do the task. Cekura Bench tests your deployed voice agent on real audio calls, three runs each. See how the two compare.