Perplexity open-sources WANDR benchmark for AI agent research capabilities

COVERAGE [7]

1. X — Perplexity TIER_1 English(EN) · perplexity_ai · 2026-07-14 18:35

   Soft scores give partial credit. Hard scores require every component for a member to be complete and correct.

   Soft scores give partial credit. Hard scores require every component for a member to be complete and correct. The benchmark tasks and evaluation harness are available at: https://t.co/o2HX0a4RTQ

2. X — Perplexity TIER_1 English(EN) · perplexity_ai · 2026-07-14 18:35

   Instead of grading against a gold solution, WANDR re-fetches every cited page and checks each claim against the underlying evidence.

   Instead of grading against a gold solution, WANDR re-fetches every cited page and checks each claim against the underlying evidence. This allows certain tasks to contain time-varying facts. https://t.co/2sDRQrSoSx

3. X — Perplexity TIER_1 English(EN) · perplexity_ai · 2026-07-14 18:35

   WANDR is constructed from de-identified production use cases covering people’s day-to-day research tasks like competitive research, due diligence, literature re

   WANDR is constructed from de-identified production use cases covering people’s day-to-day research tasks like competitive research, due diligence, literature review, market analysis, product comparison, talent sourcing, and more. https://t.co/y6dClkm2a6

4. X — Perplexity TIER_1 English(EN) · perplexity_ai · 2026-07-14 18:34

   WANDR tests how well agents discover large sets of entities and verify specific facts about each one. It provides a dense, interpretable eval signal that reveal

   WANDR tests how well agents discover large sets of entities and verify specific facts about each one. It provides a dense, interpretable eval signal that reveals whether an agent fails, and where. The pipeline also doubles as a semi-automated factory for training data. https://t…

5. X — Perplexity TIER_1 English(EN) · perplexity_ai · 2026-07-14 18:34

   WANDR represents these requirements as hierarchical, independently verifiable records.

   WANDR represents these requirements as hierarchical, independently verifiable records. It consists of 500 research tasks that require 170,495 source-backed records across three tiers of difficulty. https://t.co/OX5e1zuGbe

6. X — Perplexity TIER_1 English(EN) · perplexity_ai · 2026-07-14 18:34

   Wide-and-deep research requires two capabilities.

   Wide-and-deep research requires two capabilities. 1) Agents must search broadly enough to find all qualifying entities 2) Agents must investigate deep enough to support every claim with evidence. https://t.co/l9GE3z3FvW

7. X — Perplexity TIER_1 English(EN) · perplexity_ai · 2026-07-14 18:34

   We’re open sourcing WANDR.

   We’re open sourcing WANDR. WANDR is an internal benchmark we built and used for building deep and wide research capabilities inside Perplexity Computer. https://t.co/gp2BWjFK4d