K12-Bench - Prerequisite Reasoning: leaderboard
The Prerequisite Reasoning task of K12-Bench (2026), built from the K12-KGraph knowledge graph of Chinese K-12 mathematics, physics, chemistry and biology textbooks, as four-option multi-select questions answered zero-shot: identify the prerequisite closure and the direct successors of a concept or skill. Exact match in percent (every option must be chosen correctly); higher is better.
Metric: Prerequisite Reasoning exact match (%, all-or-nothing over the selected option set) on four-option multi-select questions generated from the K12-KGraph curriculum knowledge graph of Chinese K-12 mathematics, physics, chemistry and biology textbooks; zero-shot, answer-only, temperature 0: prerequisite closure and direct successors of a concept or skill (5,417 items); higher is better. Source: arxiv.org. Saturation forecast: Around May 2027. 10 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Flash | 34.8 | #175 |
| 2 | Gemini 2.5 Flash | 29.7 | #463 |
| 3 | Gemma 4 31B (IT) | 28.3 | #350 |
| 4 | Ministral-3-14B-Instruct | 18.4 | #1139 |
| 5 | GPT-5.2 | 17.9 | #185 |
| 6 | Qwen 3 32B | 16.9 | #972 |
| 7 | GLM-4.7 Flash | 13.4 | #1349 |
| 8 | GPT-5 Mini | 9.9 | #364 |
| 9 | GPT-4o | 9.3 | #751 |
| 10 | Llama 3 8B Instruct | 4.3 | #4852 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=k12-bench-prerequisite-reasoning · How It Works · Data refreshed daily, snapshot 2026-11-21.