SLATE Benchmark Evaluates Long-Horizon API Selection at Scale
November 4, 2026
SLATE introduces a large-scale benchmark for testing LLM agents on long-horizon tasks involving a ~1,000-tool e-commerce library. The framework moves beyond single-step tool use to evaluate hierarchical, tool-grounded procedures in realistic enterprise environments.
HOW THIS AFFECTS YOU
●
builderUse this benchmark to stress-test agents intended for complex, multi-step enterprise workflows.
●
founderThis identifies a massive market opportunity for agents capable of managing large-scale, internal software stacks.