ELBench - Basic Education: leaderboard
The basic-education module of ELBench (Jiang et al., 2026): 45 teaching scenarios from the ELMES collection covering knowledge-point explanation, question generation, lesson planning and multi-turn tutoring of a simulated student. Mean rubric score in percent from a Qwen3.6 judge, higher is better.
Metric: Module score (%; mean rubric score on 45 ELMES teaching scenarios: knowledge-point explanation, question generation, lesson planning and multi-turn tutoring of a simulated student; zero-shot, temperature 0; open-ended items scored by a Qwen3.6 rubric judge, majority of 9 calls). Source: arxiv.org. Saturation forecast: Estimated already saturated. 9 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3.5 Flash | 94.6 | #101 |
| 2 | GPT-5.4 | 94.4 | #123 |
| 3 | Claude Opus 4.8 | 92.7 | #56 |
| 4 | Seed 2.0 Pro | 89.7 | #211 |
| 5 | DeepSeek V4 Pro | 88.5 | #188 |
| 6 | DeepSeek V4 Flash | 84.9 | #258 |
| 7 | GLM-5.1 | 79.2 | #179 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=elbench-basic-education · How It Works · Data refreshed daily, snapshot 2026-11-13.