Computer Use Evals Benchmarks Deep Research

WANDR

A benchmark for high-volume, evidence-heavy knowledge work

WANDR (Wide ANd Deep Research) is Perplexity AI's open benchmark and evaluation harness for wide-and-deep research agents. It contains 500 realistic data-collection tasks requiring broad discovery, evidence-backed answer synthesis, and independently verifiable source-backed records.

Recent stories

1 linked story

releasePRIMARY2026-07-14

Perplexity releases WANDR benchmark with 500 deep-research tasks

Perplexity released WANDR, its internal benchmark for deep and wide research in Computer. The dataset has 500 tasks, 170,495 source-backed records and production-derived use cases.