Which AI best supports the people who support students?
GRADE is an open benchmark that measures how well AI systems understand educational data, interpret program trends, and recommend actions — the way a program director, site coordinator, or district leader actually needs.
Why this matters
New AI tools arrive in education every month, each promising to lighten your load. The hard question isn't whether AI is impressive — it's whether it actually helps you do right by your students. Can it explain why attendance dropped at Site C last month? Can it spot which tutors need support before they burn out? Can it help you make the case to a funder with real numbers?
GRADE answers that question safely. Every AI system here is tested on fully synthetic program data — no real student records, ever — and judged by educators on the situations you actually face. So before a tool gets anywhere near your students or your data, you know whether it understands your work, where it falls short, and whether it has earned a place in your program.
How it works
We test AI across six capability tiers, from basic data retrieval to nuanced recommendations. Tiers 1 through 4 are scored automatically. Tiers 5 and 6 — Interpretation and Recommendation — are scored by educators like you through side-by-side voting.
Your expertise shapes the rankings. When you vote in the Arena, you're telling the AI community what "helpful" actually looks like in education. The more educators participate, the more meaningful the results.
Retrieval
Can the AI find the right data when asked a direct question?
Computation
Can the AI calculate meaningful metrics from raw program data?
Temporal Reasoning
Can the AI handle time-aware queries across academic terms and years?
Ambiguity Handling
Does the AI know when to answer versus when to ask for clarification?
Interpretation
Can the AI draw meaningful, contextual conclusions from educational data?
Recommendation
Can the AI suggest actionable, educationally sound next steps?