Evaluator last updated September 23, 2025.
Overview
The Grade Level Appropriateness evaluator assesses whether AI-generated text is suitable for independent reading at a specified grade band.
The evaluator considers:
- Flesch-Kincaid grade level
- Word count
- Text structure – Organization complexity, connections between ideas, text features
- Language features – Vocabulary, sentence complexity, figurative or abstract
- Purpose – Explicitly vs. not explicitly stated, concrete vs. abstract
- Knowledge demands – Discipline-specific knowledge, references, allusions
- Student background knowledge – What students at a given grade level would already know
At a glance
| |
|---|
| Input type | Informational text |
| Supported grades | K–12 |
| Rubric | SAP ↗‘s Qualitative Text Complexity Rubric for Informational Text ↗ |
| |
|---|
| Model used | Gemini-2.5-pro (gemini-2.5-pro-preview-06-05) |
| Temperature | 0.25 |
Getting started
Follow the Quickstart to start using this evaluator:
| Input | Description | Required |
|---|
| Target grade level | Enables grade context evaluation | Yes |
| Text type | Informational text
Optimal length < 1,200 words | Yes |
Output
| Field | Description |
|---|
| Grade | Grade band where a student can read the text independently |
| Reasoning | Reasoning for the output grade level (quantitative score, qualitative features, assumed background knowledge) |
| Alternative grade | Text is appropriate for supported reading (e.g., read-aloud) |
| Scaffolding needed | Supports needed for supported reading (e.g., pre-teaching of vocabulary) |
Interpreting results
| Output | How to use |
|---|
| Grade + Reasoning | Evaluate and improve the complexity of your AI-generated content
Example: Validate that your LLM prompts produce grade-appropriate content; aggregate reasoning across runs to diagnose and fix systemic complexity issues |
| Alternative grade + Scaffolding needed | Adapt content for a wider range of learners.
Example: Make scaffolding suggestions (e.g., vocabulary pre-teaching, read-aloud) to help educators adapt content for mixed classrooms |
Accuracy and validation
This evaluator is provided as Early access. Comprehensive accuracy measures
are not yet available. Validation testing is ongoing.
| Metric | Description | Result |
|---|
| Overall accuracy | How accurately the evaluator determines grade level appropriateness compared to expert annotations. | 81% (70 correct out of 86 texts) |
| Baseline comparison | How the evaluator’s accuracy compares to a simple, unrefined grade-level appropriateness prompt. | 58% more accurate than a naive LLM baseline |
Evaluator release history
| Date | Changed |
|---|
| September 23, 2025 | First release |
