Benchmark
WritingBench
Evaluate writing against requirements specific to each assignment.
Published leaderboard
Explore scores and evaluation settings on the publisher’s leaderboard.
Open leaderboardCoverage
The revised set has 1,000 tasks in 100 subdomains across academic/engineering, finance/business, politics/law, literature/art, education, and advertising/marketing writing.
Methodology
Five instance-specific criteria score each response. The verified board uses Claude-Sonnet-4-5 and displays scores scaled to 100.