hyper-τ-bench

Product

Open long-horizon evaluation for measuring how well developer agents can construct customer-service agents from incomplete business records.

Launched: Sep 8, 2026
Models: —
Modality: —
Open source: Yes
API available: —
Releases: 1

Releases

Sep 8, 2026Sierra open-sources hyper-τ-benchhyper-τ-benchBenchmark result

Sierra released hyper-τ-bench, a long-horizon evaluation in which developer agents recover requirements, build customer-service agents in sandboxes and are scored on held-out tasks.

Sierra Benchmark resultSierra hyper-τ-bench