🛰️ Daily AI Frontier
‹ back to 2026-09-24

StudentBench: AI and human tutoring yield equivalent GRE learning gains

Research AI Tutoring

Ranking

Overall 82
Content 95
Popularity 50

Observed public metrics from 1 member.

Representative image for StudentBench: AI and human tutoring yield equivalent GRE learning gains

Merged summary

TL;DR - StudentBench finds that LLM tutors can deliver GRE learning gains statistically equivalent to expert human tutoring, with the strongest AI tutor achieving this at 918 times lower cost. The benchmark and public platform provide large-scale evaluations of lesson planning, practice generation, conversational pedagogy, cost, and engagement.

  • The study evaluated 2,383 participants receiving AI tutoring, human tutoring, or no tutoring across Quantitative and Verbal GRE questions.
  • AI tutoring was statistically equivalent to expert human tutoring overall; the best AI tutor also outperformed the human tutor on average in five of seven GRE domains.
  • Expert tutors conducted 2,028 pairwise rubric evaluations of LLM-generated lesson plans and practice problems.
  • In Quantitative GRE sessions, faster responses correlated with more student messages, which correlated with more correct practice and larger learning gains.

Sources (1)

StudentBench: AI and human tutoring yield equivalent GRE learning gains

arXiv cs.AI Curtis Northcutt, Inaara Hasmani, Kevin Feng, Trevor Khangi, Andreas Plesner, Jonas Mueller 2026-09-23 arXiv:2609.28470
Public signals Hugging Face upvotes 3
Providers: Hugging Face · Upvotes 3 OpenAlex · N/A Publisher · N/A Semantic Scholar · N/A X · N/A Fetched 2026-09-25 14:16:23.950531 UTC

TL;DR - StudentBench finds that LLM tutors can deliver GRE learning gains statistically equivalent to expert human tutoring, with the strongest AI tutor achieving this at 918 times lower cost. The benchmark and public platform provide large-scale evaluations of lesson planning, practice generation, conversational pedagogy, cost, and engagement.

  • The study evaluated 2,383 participants receiving AI tutoring, human tutoring, or no tutoring across Quantitative and Verbal GRE questions.
  • AI tutoring was statistically equivalent to expert human tutoring overall; the best AI tutor also outperformed the human tutor on average in five of seven GRE domains.
  • Expert tutors conducted 2,028 pairwise rubric evaluations of LLM-generated lesson plans and practice problems.
  • In Quantitative GRE sessions, faster responses correlated with more student messages, which correlated with more correct practice and larger learning gains.
item →