structured output benchmark
IF Leaderboard
IF is an Instruction-Following Evaluation (IFEval) benchmark for large language models that uses around 500 prompts with verifiable instructions.
Updated Sep 5, 2026
Models2
Model coverage2
MetricScore
EvidenceB
IF Ranking
Higher score ranks better on this benchmark.
Sort by
Rank | Model | Score | Percentile | Participants | Evidence | Evaluated |
|---|
| Rank01 | ModelMA | Score84.78% | Percentile100.00% | Participants2 | EvidenceC | Evaluated |
| Rank02 | ModelMI | Score72.00% | Percentile0.00% | Participants2 | EvidenceC | Evaluated |
IF Highlights
The leading models and scores on this benchmark.
IF Score Distribution
A closer view of the leading scores on this benchmark.
The Top AI Models for IF
The first five results on this benchmark, with official price and output speed added where the model identity can be matched.
What is IF?
What IF measures and how its scores work.
IF is an Instruction-Following Evaluation (IFEval) benchmark focused on 25 types of instructions, with prompts containing one or more verifiable constraints.
IF measures Score, reported as a ratio.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- IF
- Modality
- text
- Primary category
- structured output
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- overall
LLM Stats. Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about IF.
Which model scores highest on IF?
Mistral Small 3.2 24B Instruct is currently ranked first with 84.78%.
What are the top three models on IF?
The current leaders are Mistral Small 3.2 24B Instruct (84.78%) and MiniMax M2 (72.00%).
Which IF model has the lowest official input price?
MiniMax M2 has the lowest matched official input price at $0.30 input / $1.2 output per 1M tokens.
Which models are fastest among IF results?
The fastest matched records are MiniMax M2 (70.00 tok/s via MiniMax).
Does the highest score result prove overall model quality?
No. This benchmark measures one defined capability or task. The overall LLMBoard score uses a separate aggregation across eligible benchmark evidence.