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 Directory

Scoring & Data

Scoring & Data
1281 models767 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 Benchmarks
llmboard.aiCopyright 2026 llmboard.ai

instruction language context benchmark

LiveBench instruction (2026-06-25) Leaderboard

All tasks required; arithmetic mean: paraphrase, simplify, story_generation, summarize

Updated Sep 25, 2026

Models41
Model coverage41
Metricinstruction
EvidenceB

On this page

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

LiveBench instruction (2026-06-25) Ranking

Higher instruction ranks better on this benchmark.

30 of 41 rows
Columns

Show columns

Sort by
Rank
Model
instruction
Percentile
Participants
Evidence
Evaluated
Rank01ModelGOGemini 3.8 FlashGoogleinstruction81.41 scorePercentile100.00%Participants41EvidenceBEvaluatedN/A
Rank02ModelGOGemini 3.7 FlashGoogleinstruction79.93 scorePercentile97.50%Participants41EvidenceBEvaluatedN/A
Rank03ModelGOGemini 3.1 ProGoogleinstruction79.10 scorePercentile95.00%Participants41EvidenceBEvaluatedN/A
Rank04ModelMEMuse Spark 1.3Metainstruction78.00 scorePercentile92.50%Participants41EvidenceBEvaluatedN/A
Rank05ModelACQwen3.8-Flash-NextAlibaba Cloud / Qwen Teaminstruction77.11 scorePercentile90.00%Participants41EvidenceBEvaluatedN/A
Rank06ModelGOGemini 3.5 FlashGoogleinstruction75.60 scorePercentile87.50%Participants41EvidenceBEvaluatedN/A
Rank07ModelOPGPT-6 AstraOpenAIinstruction75.58 scorePercentile85.00%Participants41EvidenceBEvaluatedN/A
Rank08ModelGOGemini 3.6 FlashGoogleinstruction75.37 scorePercentile82.50%Participants41EvidenceBEvaluatedN/A
Rank09ModelXAGrok 4.7xAIinstruction75.28 scorePercentile80.00%Participants41EvidenceBEvaluatedN/A
Rank10ModelMEMuse Spark 1.2Metainstruction74.33 scorePercentile77.50%Participants41EvidenceBEvaluatedN/A
Rank11ModelACQwen3.8 MaxAlibaba Cloud / Qwen Teaminstruction74.08 scorePercentile75.00%Participants41EvidenceBEvaluatedN/A
Rank12ModelACQwen3.7 MaxAlibaba Cloud / Qwen Teaminstruction74.04 scorePercentile72.50%Participants41EvidenceBEvaluatedN/A
Rank13ModelNVNemotron 3 Ultra (550B A55B)NVIDIAinstruction73.45 scorePercentile70.00%Participants41EvidenceBEvaluatedN/A
Rank14ModelACQwen3.8-27BAlibaba Cloud / Qwen Teaminstruction72.66 scorePercentile67.50%Participants41EvidenceBEvaluatedN/A
Rank15ModelXAGrok 4.6xAIinstruction71.87 scorePercentile65.00%Participants41EvidenceBEvaluatedN/A
Rank16ModelOPGPT-5.6 SolOpenAIinstruction71.85 scorePercentile62.50%Participants41EvidenceBEvaluatedN/A
Rank17ModelXAGrok 4.5xAIinstruction71.53 scorePercentile60.00%Participants41EvidenceBEvaluatedN/A
Rank18ModelMAKimi K3Moonshot AIinstruction71.36 scorePercentile57.50%Participants41EvidenceBEvaluatedN/A
Rank19ModelDEDeepSeek-V4-Flash-Vision-ExpDeepSeekinstruction70.96 scorePercentile55.00%Participants41EvidenceBEvaluatedN/A
Rank20ModelOPGPT-5.5OpenAIinstruction70.73 scorePercentile52.50%Participants41EvidenceBEvaluatedN/A
Rank21ModelOPGPT-5.4OpenAIinstruction70.22 scorePercentile50.00%Participants41EvidenceBEvaluatedN/A
Rank22ModelTMInklingThinking Machines Labinstruction70.10 scorePercentile47.50%Participants41EvidenceBEvaluatedN/A
Rank23ModelMEMuse Spark 1.1Metainstruction69.64 scorePercentile45.00%Participants41EvidenceBEvaluatedN/A
Rank24ModelZAGLM-5.3Zhipu AIinstruction69.30 scorePercentile42.50%Participants41EvidenceBEvaluatedN/A
Rank25ModelDEDeepSeek-V4-Pro-0813DeepSeekinstruction67.70 scorePercentile40.00%Participants41EvidenceBEvaluatedN/A
Rank26ModelGOGemini 3.5 Flash-LiteGoogleinstruction67.24 scorePercentile37.50%Participants41EvidenceBEvaluatedN/A
Rank27ModelOPGPT-5.4 nanoOpenAIinstruction67.20 scorePercentile35.00%Participants41EvidenceBEvaluatedN/A
Rank28ModelOPGPT-5.2 CodexOpenAIinstruction66.45 scorePercentile32.50%Participants41EvidenceBEvaluatedN/A
Rank29ModelDEDeepSeek-V4-Flash-0731DeepSeekinstruction65.52 scorePercentile30.00%Participants41EvidenceBEvaluatedN/A
Rank30ModelOPGPT-5.6 TerraOpenAIinstruction64.62 scorePercentile27.50%Participants41EvidenceBEvaluatedN/A

LiveBench instruction (2026-06-25) Highlights

The leading models and scores on this benchmark.

LiveBench instruction (2026-06-25) Score Distribution

A closer view of the leading scores on this benchmark.

LiveBench instruction (2026-06-25)

The Top AI Models for LiveBench instruction (2026-06-25)

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

What is LiveBench instruction (2026-06-25)?

What LiveBench instruction (2026-06-25) measures and how its scores work.

All tasks required; arithmetic mean: paraphrase, simplify, story_generation, summarize

Scores are shown in score. This benchmark is verified and has an evidence level of B.

Family
livebench-if
Modality
text
Primary category
instruction language context
Score direction
higher
LLMBoard eligible
Yes
Evaluation key
overall

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

FAQ

Common questions about LiveBench instruction (2026-06-25).

Which model scores highest on LiveBench instruction (2026-06-25)?

Gemini 3.8 Flash is currently ranked first with 81.41 score.

What are the top three models on LiveBench instruction (2026-06-25)?

The current leaders are Gemini 3.8 Flash (81.41 score), Gemini 3.7 Flash (79.93 score), and Gemini 3.1 Pro (79.10 score).

Which LiveBench instruction (2026-06-25) model has the lowest official input price?

DeepSeek-V4-Flash-Vision-Exp has the lowest matched official input price at $0.15 input / $0.60 output per 1M tokens.

Which models are fastest among LiveBench instruction (2026-06-25) results?

The fastest matched records are Gemini 3.6 Flash (188.24 tok/s via Google), Gemini 3.5 Flash-Lite (185.11 tok/s via Google), and Gemini 3.8 Flash (155.60 tok/s via Google).

Does the highest instruction 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.

What does LiveBench instruction (2026-06-25) measure?

All tasks required; arithmetic mean: paraphrase, simplify, story_generation, summarize

Is a higher instruction better?

Yes. Higher values rank better for this benchmark.

How many models are compared?

41 model results are currently shown.

Does this benchmark affect the overall score?

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

Rank #1Gemini 3.8 Flash81.41 score
Rank #2Gemini 3.7 Flash79.93 score
Rank #3Gemini 3.1 Pro79.10 score
Rank #4Muse Spark 1.378.00 score

Ranking basisThis livebench instruction (2026-06-25) AI model leaderboard uses descending instruction in the benchmark original unit. The leaderboard ranking keeps matched price and speed data separate from benchmark evidence.

  1. 01
    GO
    Gemini 3.8 FlashGoogle
    instruction
    81.41 score
    Price
    $0.75 input / $3.8 output per 1M tokens
    Speed
    Up to 155.60 tok/s via Google

    Strengths

    • Ranks #1 of 41 compared models
    • 100th percentile on this benchmark
    • B evidence result

    Considerations

    • This result measures LiveBench instruction (2026-06-25), not total model capability
  2. 02
    GO
    Gemini 3.7 FlashGoogle
    instruction
    79.93 score
    Price
    $0.75 input / $3.8 output per 1M tokens
    Speed
    Up to 42.46 tok/s via Google

    Strengths

    • Ranks #2 of 41 compared models
    • 98th percentile on this benchmark
    • B evidence result

    Considerations

    • This result measures LiveBench instruction (2026-06-25), not total model capability
  3. 03
    GO
    Google
    instruction
    79.10 score
    Price
    $2.0 input / $12 output per 1M tokens
    Speed
    Up to 90.00 tok/s via Google

    Strengths

    • Ranks #3 of 41 compared models
    • 95th percentile on this benchmark
    • B evidence result

    Considerations

    • This result measures LiveBench instruction (2026-06-25), not total model capability
  4. 04
    ME
    Meta
    instruction
    78.00 score
    Price
    $1.3 input / $4.3 output per 1M tokens
    Speed
    Up to 4.64 tok/s via Meta Model API

    Strengths

    • Ranks #4 of 41 compared models
    • 93th percentile on this benchmark
    • B evidence result

    Considerations

    • This result measures LiveBench instruction (2026-06-25), not total model capability
  5. 05
    AC
    Alibaba Cloud / Qwen Team
    instruction
    77.11 score

    Strengths

    • Ranks #5 of 41 compared models
    • 90th percentile on this benchmark
    • B evidence result

    Considerations

    • This result measures LiveBench instruction (2026-06-25), not total model capability

Selection summary

Best AI Models for LiveBench instruction (2026-06-25)

Gemini 3.8 Flash currently leads LiveBench instruction (2026-06-25) with 81.41 score. 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.

Gemini 3.1 Pro
Muse Spark 1.3
Qwen3.8-Flash-Next
Benchmark rank #1Gemini 3.8 Flash81.41 score · $0.75 input / $3.8 output per 1M tokens
Benchmark rank #2Gemini 3.7 Flash79.93 score · $0.75 input / $3.8 output per 1M tokens
Benchmark rank #3Gemini 3.1 Pro79.10 score · $2.0 input / $12 output per 1M tokens