# IFEval

> IFEval (Instruction-Following Eval) is a benchmark of prompts containing instructions whose compliance can be checked automatically, such as "write at least 300 words" or "do not use any commas", used to measure instruction following without a human or model judge.

- Identifier: PTL-0208
- Category: Benchmarks
- Canonical URL: https://protologue.com/t/ifeval/
- Also known as: Instruction-Following Eval
- Introduced: 2023

## Description

Verifiable instructions make scoring objective and cheap, which is why IFEval is a common headline score for instruction-tuned models. It measures adherence to format and content constraints, not answer quality.

## Broader terms

- [Benchmark](https://protologue.com/t/benchmark/)

## Related terms

- [Instruction Following](https://protologue.com/t/instruction-following/)
- [MT-Bench](https://protologue.com/t/mt-bench/)

## Sources

- Zhou et al. (2023). Instruction-Following Evaluation for Large Language Models. https://arxiv.org/abs/2311.07911

## Cite this entry

Protologue. (2026). IFEval. In Protologue: A Taxonomy of Prompting and LLM Techniques (v1.4.0, PTL-0208). https://protologue.com/t/ifeval/

License: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/)