llmboard.aiAI model intelligence
Home

Model Rankings

OverallOpen ModelsAgentCodingReasoningMathKnowledgeInstruction FollowingTextVision
Image GenerationImage Editing
Video GenerationImage to VideoVideo Editing
Text to SpeechSpeech to Text
Embeddings

Efficiency

Chat Token PricingImage PricingVideo PricingAudio Pricing
Chat Speed & LatencyProvider Reliability

Benchmarks

GPQAMMLU-ProAIME 2025SWE-Bench VerifiedMMLUHumanity's Last ExamLiveCodeBenchMATHHumanEvalMMMU-Pro
All Benchmarks

Tools

Model DirectoryCompare Models

Scoring & Data

Scoring & Data
393 models668 benchmarks

Leaderboard Center

Overall RankingCodingCore BenchmarksPrice & ValueRuntime Performance

Modalities

All ModelsImage GenerationImage EditingVideo GenerationImage-to-VideoVideo EditingText-to-SpeechSpeech-to-TextEmbeddings

Data & Methods

Scoring MethodAll BenchmarksReasoningMath

Vendors

All VendorsOpenAIAnthropicGoogle
llmboard.aiCopyright 2026 llmboard.ai

structured output benchmark

IFEval Leaderboard

Instruction-Following Evaluation (IFEval) benchmark for large language models, focusing on verifiable instructions with 25 types of instructions and around 500 prompts containing one or more verifiable constraints

Updated Aug 17, 2026

Models67
Model coverage67
MetricScore
EvidenceB

On this page

  • Ranking
  • Highlights
  • Distribution
  • Top models
  • About
  • FAQ

IFEval Ranking

Higher score ranks better on this benchmark.

30 of 67 rows
Columns

Show columns

Sort by
Rank
Model
Score
Percentile
Participants
Evidence
Evaluated
Rank01ModelACQwen3.5-27BAlibaba Cloud / Qwen TeamScore95.0%Percentile100.0%Participants67EvidenceCEvaluatedAug 17, 2026
Rank02ModelACQwen3.7-PlusAlibaba Cloud / Qwen TeamScore94.6%Percentile98.5%Participants67EvidenceCEvaluatedAug 17, 2026
Rank03ModelACQwen3.6 PlusAlibaba Cloud / Qwen TeamScore94.3%Percentile97.0%Participants67EvidenceCEvaluatedAug 17, 2026
Rank04ModelACQwen3.7 MaxAlibaba Cloud / Qwen TeamScore94.3%Percentile95.5%Participants67EvidenceCEvaluatedAug 17, 2026
Rank05ModelOPo3-miniOpenAIScore93.9%Percentile93.9%Participants67EvidenceCEvaluatedAug 17, 2026
Rank06ModelACQwen3.5-122B-A10BAlibaba Cloud / Qwen TeamScore93.4%Percentile92.4%Participants67EvidenceCEvaluatedAug 17, 2026
Rank07ModelANClaude 3.7 SonnetAnthropicScore93.2%Percentile90.9%Participants67EvidenceCEvaluatedAug 17, 2026
Rank08ModelACQwen3.5-397B-A17BAlibaba Cloud / Qwen TeamScore92.6%Percentile89.4%Participants67EvidenceCEvaluatedAug 17, 2026
Rank09ModelMELlama 3.3 70B InstructMetaScore92.1%Percentile87.9%Participants67EvidenceCEvaluatedAug 17, 2026
Rank10ModelAMNova ProAmazonScore92.1%Percentile86.4%Participants67EvidenceCEvaluatedAug 17, 2026
Rank11ModelACQwen3.5-35B-A3BAlibaba Cloud / Qwen TeamScore91.9%Percentile84.8%Participants67EvidenceCEvaluatedAug 17, 2026
Rank12ModelACQwen3.5-9BAlibaba Cloud / Qwen TeamScore91.5%Percentile83.3%Participants67EvidenceCEvaluatedAug 17, 2026
Rank13ModelGOGemma 3 27BGoogleScore90.4%Percentile81.8%Participants67EvidenceCEvaluatedAug 17, 2026
Rank14ModelNVNemotron Nano 9B v2NVIDIAScore90.3%Percentile80.3%Participants67EvidenceCEvaluatedAug 17, 2026
Rank15ModelGOGemma 3 4BGoogleScore90.2%Percentile78.8%Participants67EvidenceCEvaluatedAug 17, 2026
Rank16ModelMAKimi K2 InstructMoonshot AIScore89.8%Percentile77.3%Participants67EvidenceCEvaluatedAug 17, 2026
Rank17ModelMAKimi K2-Instruct-0905Moonshot AIScore89.8%Percentile75.8%Participants67EvidenceCEvaluatedAug 17, 2026
Rank18ModelACQwen3.5-4BAlibaba Cloud / Qwen TeamScore89.8%Percentile74.2%Participants67EvidenceCEvaluatedAug 17, 2026
Rank19ModelAMNova LiteAmazonScore89.7%Percentile72.7%Participants67EvidenceCEvaluatedAug 17, 2026
Rank20ModelMELongCat-Flash-ChatMeituanScore89.6%Percentile71.2%Participants67EvidenceCEvaluatedAug 17, 2026
Rank21ModelNVLlama 3.1 Nemotron Ultra 253B v1NVIDIAScore89.5%Percentile69.7%Participants67EvidenceCEvaluatedAug 17, 2026
Rank22ModelGOGemma 3 12BGoogleScore88.9%Percentile68.2%Participants67EvidenceCEvaluatedAug 17, 2026
Rank23ModelACQwen3-Next-80B-A3B-ThinkingAlibaba Cloud / Qwen TeamScore88.9%Percentile66.7%Participants67EvidenceCEvaluatedAug 17, 2026
Rank24ModelACQwen3-235B-A22B-Instruct-2507Alibaba Cloud / Qwen TeamScore88.7%Percentile65.2%Participants67EvidenceCEvaluatedAug 17, 2026
Rank25ModelMELlama 3.1 405B InstructMetaScore88.6%Percentile63.6%Participants67EvidenceCEvaluatedAug 17, 2026
Rank26ModelOPGPT-4.5OpenAIScore88.2%Percentile62.1%Participants67EvidenceCEvaluatedAug 17, 2026
Rank27ModelACQwen3 VL 235B A22B ThinkingAlibaba Cloud / Qwen TeamScore88.2%Percentile60.6%Participants67EvidenceCEvaluatedAug 17, 2026
Rank28ModelACQwen3-235B-A22B-Thinking-2507Alibaba Cloud / Qwen TeamScore87.8%Percentile59.1%Participants67EvidenceCEvaluatedAug 17, 2026
Rank29ModelACQwen3 VL 235B A22B InstructAlibaba Cloud / Qwen TeamScore87.8%Percentile57.6%Participants67EvidenceCEvaluatedAug 17, 2026
Rank30ModelACQwen3 VL 32B ThinkingAlibaba Cloud / Qwen TeamScore87.8%Percentile56.1%Participants67EvidenceCEvaluatedAug 17, 2026
Rank31ModelACQwen3-Next-80B-A3B-InstructAlibaba Cloud / Qwen TeamScore87.6%Percentile54.5%Participants67EvidenceCEvaluatedAug 17, 2026
Rank32ModelMELlama 3.1 70B InstructMetaScore87.5%Percentile53.0%Participants67EvidenceCEvaluatedAug 17, 2026
Rank33ModelOPGPT-4.1OpenAIScore87.4%Percentile51.5%Participants67EvidenceCEvaluatedAug 17, 2026
Rank34ModelMAKimi-k1.5Moonshot AIScore87.2%Percentile50.0%Participants67EvidenceCEvaluatedAug 17, 2026
Rank35ModelAMNova MicroAmazonScore87.2%Percentile48.5%Participants67EvidenceCEvaluatedAug 17, 2026
Rank36ModelDEDeepSeek-V3DeepSeekScore86.1%Percentile47.0%Participants67EvidenceCEvaluatedAug 17, 2026
Rank37ModelACQwen3 VL 30B A3B InstructAlibaba Cloud / Qwen TeamScore85.8%Percentile45.5%Participants67EvidenceCEvaluatedAug 17, 2026
Rank38ModelMIPhi 4 Reasoning PlusMicrosoftScore84.9%Percentile43.9%Participants67EvidenceCEvaluatedAug 17, 2026
Rank39ModelSASarvam-105BSarvam AIScore84.8%Percentile42.4%Participants67EvidenceCEvaluatedAug 17, 2026
Rank40ModelACQwen3 VL 32B InstructAlibaba Cloud / Qwen TeamScore84.7%Percentile40.9%Participants67EvidenceCEvaluatedAug 17, 2026
Rank41ModelOPGPT-4.1 miniOpenAIScore84.1%Percentile39.4%Participants67EvidenceCEvaluatedAug 17, 2026
Rank42ModelACQwen2.5 72B InstructAlibaba Cloud / Qwen TeamScore84.1%Percentile37.9%Participants67EvidenceCEvaluatedAug 17, 2026
Rank43ModelACQwQ-32BAlibaba Cloud / Qwen TeamScore83.9%Percentile36.4%Participants67EvidenceCEvaluatedAug 17, 2026
Rank44ModelACQwen3 VL 8B InstructAlibaba Cloud / Qwen TeamScore83.7%Percentile34.9%Participants67EvidenceCEvaluatedAug 17, 2026
Rank45ModelMIPhi 4 ReasoningMicrosoftScore83.4%Percentile33.3%Participants67EvidenceCEvaluatedAug 17, 2026
Rank46ModelACQwen3 VL 8B ThinkingAlibaba Cloud / Qwen TeamScore83.2%Percentile31.8%Participants67EvidenceCEvaluatedAug 17, 2026
Rank47ModelMAMistral Small 3 24B InstructMistral AIScore82.9%Percentile30.3%Participants67EvidenceCEvaluatedAug 17, 2026
Rank48ModelACQwen3 VL 4B ThinkingAlibaba Cloud / Qwen TeamScore82.6%Percentile28.8%Participants67EvidenceCEvaluatedAug 17, 2026
Rank49ModelLALFM2.5-VL-3BLiquid AIScore82.3%Percentile27.3%Participants67EvidenceCEvaluatedAug 17, 2026
Rank50ModelACQwen3 VL 4B InstructAlibaba Cloud / Qwen TeamScore82.3%Percentile25.8%Participants67EvidenceCEvaluatedAug 17, 2026
Rank51ModelACQwen3 VL 30B A3B ThinkingAlibaba Cloud / Qwen TeamScore81.7%Percentile24.2%Participants67EvidenceCEvaluatedAug 17, 2026
Rank52ModelOPGPT-4oOpenAIScore81.0%Percentile22.7%Participants67EvidenceCEvaluatedAug 17, 2026
Rank53ModelMELlama 3.1 8B InstructMetaScore80.4%Percentile21.2%Participants67EvidenceCEvaluatedAug 17, 2026
Rank54ModelGOGemma 3 1BGoogleScore80.2%Percentile19.7%Participants67EvidenceCEvaluatedAug 17, 2026
Rank55ModelNVLlama 3.1 Nemotron Nano 8B V1NVIDIAScore79.3%Percentile18.2%Participants67EvidenceCEvaluatedAug 17, 2026
Rank56ModelACQwen3.5-2BAlibaba Cloud / Qwen TeamScore78.6%Percentile16.7%Participants67EvidenceCEvaluatedAug 17, 2026
Rank57ModelMELlama 3.2 3B InstructMetaScore77.4%Percentile15.2%Participants67EvidenceCEvaluatedAug 17, 2026
Rank58ModelOPMiniCPM-SALAOpenBMBScore76.3%Percentile13.6%Participants67EvidenceCEvaluatedAug 17, 2026
Rank59ModelCONorth Micro Vision InstructCohereScore74.9%Percentile12.1%Participants67EvidenceCEvaluatedAug 17, 2026
Rank60ModelIBGranite 3.3 8B BaseIBMScore74.8%Percentile10.6%Participants67EvidenceCEvaluatedAug 17, 2026
Rank61ModelIBGranite 3.3 8B InstructIBMScore74.8%Percentile9.1%Participants67EvidenceCEvaluatedAug 17, 2026
Rank62ModelOPGPT-4.1 nanoOpenAIScore74.5%Percentile7.6%Participants67EvidenceCEvaluatedAug 17, 2026
Rank63ModelACQwen2.5 7B InstructAlibaba Cloud / Qwen TeamScore71.2%Percentile6.1%Participants67EvidenceCEvaluatedAug 17, 2026
Rank64ModelIBIBM Granite 4.0 Tiny PreviewIBMScore63.0%Percentile4.5%Participants67EvidenceCEvaluatedAug 17, 2026
Rank65ModelMIPhi 4MicrosoftScore63.0%Percentile3.0%Participants67EvidenceCEvaluatedAug 17, 2026
Rank66ModelMAPixtral-12BMistral AIScore61.3%Percentile1.5%Participants67EvidenceCEvaluatedAug 17, 2026
Rank67ModelACQwen3.5-0.8BAlibaba Cloud / Qwen TeamScore44.0%Percentile0.0%Participants67EvidenceCEvaluatedAug 17, 2026

IFEval Highlights

The leading models and scores on this benchmark.

Rank #1Qwen3.5-27B95.0%Rank #2Qwen3.7-Plus94.6%Rank #3Qwen3.6 Plus94.3%Rank #4Qwen3.7 Max94.3%

IFEval Score Distribution

A closer view of the leading scores on this benchmark.

IFEval

The Top AI Models for IFEval

The first five results on this benchmark, with official price and output speed added where the model identity can be matched.

Ranking basisThis ifeval AI model leaderboard uses descending score in the benchmark's original unit. The leaderboard ranking keeps matched price and speed data separate from benchmark evidence.

  1. 01
    AC
    Qwen3.5-27BAlibaba Cloud / Qwen Team
    Score
    95.0%
    Price
    $0.30 input / $2.4 output per 1M tokens
    Speed
    Up to 6.7 tok/s via Novita

    Strengths

    • Ranks #1 of 67 compared models
    • 100th percentile on this benchmark
    • C evidence result

    Considerations

    • This result measures IFEval, not total model capability
  2. 02
    AC
    Qwen3.7-PlusAlibaba Cloud / Qwen Team
    Score
    94.6%
    Price
    $0.50 input / $3.0 output per 1M tokens

    Strengths

    • Ranks #2 of 67 compared models
    • 98th percentile on this benchmark
    • C evidence result

    Considerations

    • This result measures IFEval, not total model capability
  3. 03
    AC
    Qwen3.6 PlusAlibaba Cloud / Qwen Team
    Score
    94.3%
    Price
    $0.50 input / $3.0 output per 1M tokens
    Speed
    Up to 16 tok/s via Together

    Strengths

    • Ranks #3 of 67 compared models
    • 97th percentile on this benchmark
    • C evidence result

    Considerations

    • This result measures IFEval, not total model capability
  4. 04
    AC
    Qwen3.7 MaxAlibaba Cloud / Qwen Team
    Score
    94.3%
    Price
    $2.5 input / $7.5 output per 1M tokens
    Speed
    Up to 5.8 tok/s via Together

    Strengths

    • Ranks #4 of 67 compared models
    • 95th percentile on this benchmark
    • C evidence result

    Considerations

    • This result measures IFEval, not total model capability
  5. 05
    OP
    o3-miniOpenAI
    Score
    93.9%
    Price
    $1.1 input / $4.4 output per 1M tokens
    Speed
    Up to 115 tok/s via Azure

    Strengths

    • Ranks #5 of 67 compared models
    • 94th percentile on this benchmark
    • C evidence result

    Considerations

    • This result measures IFEval, not total model capability

Selection summary

Best AI Models for IFEval

Qwen3.5-27B currently leads IFEval with 95.0%. It is the top model on this specific benchmark, while the best LLM for the broader task should also be checked against other benchmarks, price and runtime.

Use this leaderboard with the supporting benchmark results and coverage details above. A leaderboard position summarizes the selected ranking signal; it does not replace workload-specific testing.

Benchmark rank #1Qwen3.5-27B95.0% · $0.30 input / $2.4 output per 1M tokensBenchmark rank #2Qwen3.7-Plus94.6% · $0.50 input / $3.0 output per 1M tokensBenchmark rank #3Qwen3.6 Plus94.3% · $0.50 input / $3.0 output per 1M tokens

What is IFEval?

What IFEval measures and how its scores work.

Instruction-Following Evaluation (IFEval) benchmark for large language models, focusing on verifiable instructions with 25 types of instructions and around 500 prompts containing one or more verifiable constraints

Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.

Family
IFEval
Modality
text
Primary category
structured output
Score direction
higher
LLMBoard eligible
Yes
Evaluation key
ifeval|llm-stats-current

Benchmark scores retain their original unit. Overall score eligibility is shown separately.

FAQ

Common questions about IFEval.

Which model scores highest on IFEval?

Qwen3.5-27B is currently ranked first with 95.0%.

What does IFEval measure?

Instruction-Following Evaluation (IFEval) benchmark for large language models, focusing on verifiable instructions with 25 types of instructions and around 500 prompts containing one or more verifiable constraints

Is a higher score better?

Yes. Higher values rank better for this benchmark.

How many models are compared?

67 model results are currently shown.

Does this benchmark affect the overall score?

Yes. This benchmark can contribute to the current LLMBoard capability score.