XpertBench - Education: leaderboard

The education tasks of XpertBench (2026), open-ended professional assignments written by domain experts and graded against typically 15 to 40 weighted expert rubric checkpoints per task. Weighted rubric score in percent on the education part of the XpertBench-Gold subset, judged by Gemini 2.5 Pro; higher is better.

Metric: Weighted rubric score (%) on the XpertBench-Gold subset (245 expert-written open-ended professional tasks, 15-40 weighted binary checkpoints each); ShotJudge: Gemini 2.5 Pro judges every checkpoint with an expert-graded GPT-5 response as one-shot exemplar; education tasks; higher is better. Source: arxiv.org. Saturation forecast: Around May 2027. 14 models tracked. Updated 7 Oct 2026.

Top models

#ModelSettingScoreOverall rank
1GPT-5.4High59.29#135
2GPT-5.2High58.43#185
3Claude Opus 4.6Thinking57.96#113
4GPT-5High56.9#178
5Doubao-2.0-pro55.35#149
6Kimi K2 (Thinking)54.4
7Gemini 3.1 Pro (Preview)52.43#99
8Gemini 3 Pro48.59#145
9Gemini 2.5 Pro47.9#276
10Gemini 3 Flash47.1#175
11Doubao-Seed-1.6Thinking45.9
12Claude Sonnet 4.5Thinking44.5#261
13DeepSeek V3.1 TerminusThinking41.9#437
14GLM-4.639.1#498

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=xpertbench-education · How It Works · Data refreshed daily, snapshot 2026-11-21.