Original research and data from DualEntry for finance and accounting teams in the AI era. Every benchmark publishes its methodology in full, and all of it is free to read.
We test the frontier models on real accounting work: classification, journal entries, AP/AR, reconciliation, reporting, and close. Grading is deterministic and each run happens in an isolated environment, repeated to report accuracy and standard deviation per category.
What 288 finance leaders actually earn across base, bonus, equity, and severance, benchmarked by company size, geography, and financing source. Produced by Benchmarkit, presented by DualEntry with The SaaS CFO and FIF Collective.
Benchmark questions run against a provisioned chart of accounts, and the compensation data comes from a survey of 288 finance leaders.
Binary pass or fail per task, repeated runs, accuracy and standard deviation reported per category.
Categories, question counts, and how the grading works are published next to the results, not just the headline.
See how DualEntry uses AI across the close, or run your own numbers with our free finance calculators.