OpenLearnLM - Content Knowledge: leaderboard
Subject knowledge part of the OpenLearnLM Benchmark (Lee et al., 2026): four-option multiple-choice questions drawn from C-Eval middle-school subjects and GPQA science, scored by exact-match accuracy in percent on the benchmark's test split; higher is better.
Metric: Accuracy (%; four-option multiple choice from C-Eval and GPQA). Source: arxiv.org. Saturation forecast: Around February 2027. 7 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Grok 4.1 Fast | 86.5 | #426 |
| 2 | Gemini 3 Pro | 82.4 | #135 |
| 3 | Kimi K2 (Thinking) | 81.4 | #435 (Kimi K2) |
| 4 | GPT-5.2 | 80.8 | #172 |
| 5 | GLM-4.7 | 80.8 | #365 |
| 6 | DeepSeek V3.2 | 74.6 | #403 |
| 7 | Claude Opus 4.5 | 66.3 | #152 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=openlearnlm-content-knowledge · How It Works · Data refreshed daily, snapshot 2026-11-12.