SciEval (K-12) - Evidence Match Rate: leaderboard
Metric: Evidence match rate (%): share of predicted evidence sentences whose best sentence-embedding cosine similarity to an expert evidence sentence exceeds 0.65, on the held-out test set of SciEval, a K-12 science instructional-material evaluation benchmark: given an NGSS lesson as page-marked PDF text and one EQuIP rubric criterion, the model returns a 0-3 score and the supporting evidence as JSON, simplified prompt, temperature 0; failed predictions are left out; higher is better. Source: arxiv.org. Saturation forecast: Around 2034. 6 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3 4B 2507 Instruct | 36.04 |
| 2 | Gemini 2.0 Flash | 18.27 |
| 3 | Llama 3.1 8B Instruct | 17.25 |
| 4 | Llama 3.2 3B Instruct | 14.97 |
| 5 | GPT-4o Mini | 8.63 |
Interactive version: theaggregate.ai/benchmark?slug=scieval-k-12-evidence-match-rate · How It Works · Data refreshed daily, snapshot 2026-10-07.