The best AI tutors exceeded human mean learning gains in Quantitative domains while human tutoring led pooled AI means in all Verbal domains¶

Claim · All claims
Evidence · 1 study · causal r2 · q2
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q2 i? The highest-mean AI tutor beat human tutoring in three of four Quantitative domains, and human mean gains remained above the pooled AI mean in all three Verbal domains. → Curtis Northcutt 2026

Evidence¶

Curtis Northcutt 2026¶

Curtis Northcutt, Inaara Hasmani, Kevin Feng, Trevor Khangi, Andreas Plesner, Jonas Mueller. (2026). StudentBench: AI and human tutoring yield equivalent GRE learning gains. https://arxiv.org/abs/2609.28470 doi:10.48550/arxiv.2609.28470

q2 · i? · causal · r2

Descriptive pattern across the seven GRE domains shown in Figure 1A, with 95% confidence intervals per condition; the text states "the best AI tutors have higher mean gains than humans in three of the four domains". Descriptive means, no test printed.

"In Quantitative, pooled AI and human mean learning gains are generally close, while the best AI tutors have higher mean gains than humans in three of the four domains. In Verbal, human mean gains are close to those of the best AI tutors, but remain above the pooled AI mean in all three domains."

Discussion¶

Related Claims¶

• Both AI and human tutoring produce larger GRE learning gains than a no-tutoring control — related
• Pooled AI tutoring produces statistically equivalent GRE learning gains to expert human tutoring — related
• In Quantitative AI sessions, faster AI replies were associated with more engagement, more correct practice, and larger learning gains; these associations were not significant in Verbal — related
• Six AI tutors individually matched human tutoring, with Gemma 4 31B equivalent at 918× lower cost per percentage point of learning gain — related
• Scenario-based tutor lessons produced a significant pooled 7.4% learning gain, driven by three lessons with established construct validity while three newer lessons showed no significant gains — related