StudentBench - Learning Gain: leaderboard
Real students take a GRE practice pre-test, are tutored by an AI model for an hour, then take a different post-test; the score is the pretest-adjusted learning gain in percentage points for students of that tutor. Each row is one model at one reasoning setting. Higher is better. Source: studentbench.org.
Metric: GRE learning gain after one hour of AI tutoring (percentage points, pretest-adjusted ANCOVA). Source: studentbench.org. Saturation forecast: Around January 2028. 11 models tracked. Updated 4 Oct 2026.
Top models
| # | Model | Setting | Score | Overall rank |
|---|---|---|---|---|
| 1 | Claude Opus 4.8 | xHigh | 16 | #64 |
| 2 | GPT-5.5 Pro | Medium | 15.7 | #40 |
| 3 | Gemini 3.1 Pro (Preview) | High | 15.4 | #101 |
| 4 | Gemini 3.5 Flash | Low | 14.3 | #108 |
| 5 | Gemini 3.7 Flash | Medium | 14.1 | #42 |
| 6 | GPT-5.5 | High | 13.5 | #54 |
| 7 | Gemma 4 31B | High | 13.1 | #389 |
| 8 | GPT-5.4 Mini | Non-reasoning | 13 | #405 |
| 9 | Claude Opus 5 | High | 12.6 | #26 |
| 10 | Kimi K2.6 | Medium | 12.2 | #192 |
| 11 | Gemini 3.6 Flash | Low | 12.1 | #79 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=studentbench-learning-gain · How It Works · Data refreshed daily, snapshot 2026-11-20.