SQuALITY: leaderboard

Question-focused summarisation of 127 Project Gutenberg short stories of 4,000 to 6,000 words, each with five questions and four human-written reference summaries per question, from NYU (2022).

Source: arxiv.org.

Interactive version: theaggregate.ai/benchmark?slug=squality · How It Works · Data refreshed daily, snapshot 2026-09-05.