Long-Horizon Augmented Workflows

Product

Evaluation for how agents detect underspecification, request clarification, and recover on extended ambiguous tasks.

Launched: Mar 9, 2026
Models: —
Modality: —
Open source: —
API available: —
Releases: 1

Releases

Jan 23, 2026Long-Horizon Augmented Workflows releasedLong-Horizon Augmented WorkflowsBenchmark result

Scale researchers released LHAW, a framework for generating underspecified long-horizon tasks and evaluating whether agents clarify ambiguity and recover performance.

Scale Labs Benchmark resultScale AI Long-Horizon Augmented Workflows