AnotherWrapper
    $249
    LLMImageVideoSpeechTranscription

    MMLU

    🏆 Leaderboard

    As of August 20, 2026, GPT-5 is #1 for MMLU at 92.5%. Ranked by MMLU, a multiple-choice knowledge exam. It does not measure prose. 87 models in this index have a published MMLU score. Methodology: MMLU (https://github.com/hendrycks/test). MMLU leaderboard with live API prices. Massive Multitask Language Understanding — tests knowledge across 57 subjects including STEM, humanities, and social sciences. Official methodology: MMLU (https://github.com/hendrycks/test).

    Updated August 20, 2026282 models33 providers
    GPT-5
    OpenAI · Proprietary
    		92.5%$11.25
    o1
    OpenAI · Proprietary
    		91.8%$75.00
    o1-preview
    OpenAI · Proprietary
    		90.8%$75.00
    Qwen3 VL 235B A22B ThinkingOSS
    Qwen · Open Source
    		90.6%$3.94
    Claude 3.5 Sonnet
    Anthropic · Proprietary
    		90.4%$18.00
    Claude 3.5 Sonnet
    Anthropic · Proprietary
    		90.4%$18.00
    Kimi K2 0905
    Moonshot AI · Proprietary
    		90.2%$3.10
    GPT-4.1
    OpenAI · Proprietary
    		90.2%$10.00
    GPT OSS 120BOSS
    OpenAI · Open Source
    		90%$0.54
    LongCat-Flash-ChatOSS
    Meituan · Open Source
    		89.7%$1.50
    Kimi K2-Instruct-0905OSS
    Moonshot AI · Open Source · via OpenRouter
    		89.5%$3.10
    Kimi K2 InstructOSS
    Moonshot AI · Open Source
    		89.5%$1.00
    MiMo-V2.5-ProOSS
    Xiaomi · Open Source
    		89.4%$1.30
    Claude Sonnet 4.6
    Anthropic · Proprietary
    		89.3%$18.00
    Gemini 2.5 Pro
    Google · Proprietary
    		88.9%$11.25
    Qwen3 VL 235B A22B InstructOSS
    Qwen · Open Source
    		88.8%$1.79
    GPT-4o
    OpenAI · Proprietary
    		88.7%$12.50
    DeepSeek-V3OSS
    DeepSeek · Open Source
    		88.5%$1.37
    Qwen3 235B A22BOSS
    Qwen · Open Source
    		87.8%$0.20
    Qwen3.5-397B-A17BOSS
    Qwen · Open Source
    		87.8%$4.20
    Qwen3 VL 30B A3B ThinkingOSS
    Qwen · Open Source
    		87.6%$1.20
    Grok-2
    xAI · Proprietary
    		87.5%$12.00
    GPT-4.1 mini
    OpenAI · Proprietary
    		87.5%$2.00
    Llama 3.1 405B InstructOSS
    Meta · Open Source
    		87.3%$1.78
    o3-mini
    OpenAI · Proprietary
    		86.9%$5.50
    Claude 3 Opus
    Anthropic · Proprietary
    		86.8%$90.00
    GPT-4 Turbo
    OpenAI · Proprietary
    		86.5%$40.00
    Qwen3 VL 32B InstructOSS
    Qwen · Open Source · via OpenRouter
    		86.4%$0.52
    GPT-4
    OpenAI · Proprietary
    		86.4%$90.00
    Llama 3.3 70B InstructOSS
    Meta · Open Source
    		86%$0.40
    Llama 3.2 90B InstructOSS
    Meta · Open Source
    		86%$0.75
    Nova Pro
    Amazon · Proprietary
    		85.9%$4.00
    Gemini 1.5 Pro
    Google · Proprietary
    		85.9%$12.50
    GPT-4o
    OpenAI · Proprietary
    		85.7%$12.50
    LongCat-Flash-LiteOSS
    Meituan · Open Source
    		85.5%$0.50
    Llama 4 MaverickOSS
    Meta · Open Source
    		85.5%$1.02
    GPT OSS 20BOSS
    OpenAI · Open Source
    		85.3%$0.25
    Qwen3 VL 8B ThinkingOSS
    Qwen · Open Source
    		85.2%$2.27
    o1-mini
    OpenAI · Proprietary
    		85.2%$15.00
    Qwen3 VL 30B A3B InstructOSS
    Qwen · Open Source
    		85%$0.90
    Phi 4OSS
    Microsoft · Open Source
    		84.8%$0.21
    GLM-4.5OSS
    Z AI · Open Source
    		84.6%$2.80
    GLM-4.7OSS
    Z AI · Open Source
    		84.3%$2.80
    Qwen2.5 72B InstructOSS
    Qwen · Open Source
    		84.2%$0.75
    Mistral Large 2OSS
    Mistral · Open Source · via Mistral AI
    		84%$8.00
    Llama 3.1 70B InstructOSS
    Meta · Open Source
    		83.6%$0.40
    Grok-3 Mini
    xAI · Proprietary
    		82.8%$0.80
    GPT-4o mini
    OpenAI · Proprietary
    		82%$0.75
    Qwen3 VL 4B ThinkingOSS
    Qwen · Open Source
    		81.5%$1.10
    Jamba 1.5 LargeOSS
    AI21 Labs · Open Source
    		81.2%$10.00
    Showing 1–50 of 282 models

    Next step

    You found the model. Now ship the product.

    Auth, billing, and the API layer are already decided. Fork 8 finished AI apps, or have us build the first version with you.

    Get the stack — $249Have us build it with you

    The index

    All Large Language Models

    282 models across 33 providers. Search or jump to a lab — every model page stays linked here.

    OpenAI51Anthropic22Google25Cursor2xAI15DeepSeek16MiniMax6Qwen38AI21 Labs2Amazon5Baidu1ByteDance2Cohere3IBM1Inception1inclusionAI1LG AI Research1Meituan4Meta12Microsoft5Mistral23Moonshot AI9Nous Research1NVIDIA5Poolside2Sakana AI1StepFun1Tencent1Thinking Machines1Unisound1Upstage1Xiaomi3Z AI20

    Baidu

    1 models
    ERNIE 4.5

    Inception

    1 models
    Mercury 2

    inclusionAI

    1 models
    Ling-3.0-flash

    LG AI Research

    1 models
    K-EXAONE-236B-A23B

    Nous Research

    1 models
    Hermes 3 70B

    Sakana AI

    1 models
    Sakana Namazu

    StepFun

    1 models
    Step-3.5-Flash

    Tencent

    1 models
    Hy3

    Thinking Machines

    1 models
    Inkling-Small

    Unisound

    1 models
    U2

    Upstage

    1 models
    Solar Pro 4

    Which model leads MMLU right now?

    As of August 20, 2026, GPT-5 by OpenAI is #1 for MMLU at 92.5%. Ranked by MMLU, a multiple-choice knowledge exam. It does not measure prose. This board also tracks MMLU. Next on the same board: o1 and o1-preview. This mmlu leaderboard ranks models by MMLU. Scores come from public evals. Prices are the live API rates in the table above.

    Sources: MMLU (https://github.com/hendrycks/test); OpenAI API pricing (https://developers.openai.com/api/docs/pricing); Anthropic Claude API pricing (https://platform.claude.com/docs/en/about-claude/pricing); Gemini API pricing (https://ai.google.dev/gemini-api/docs/pricing)

    Top 8 for mmlu. Ranked by MMLU, a multiple-choice knowledge exam. It does not measure prose. Input and output are dollars per million tokens.
    RankModelMMLUInput /MOutput /M
    1GPT-592.5%$1.25$10.00
    2o191.8%$15.00$60.00
    3o1-preview90.8%$15.00$60.00
    4Qwen3 VL 235B A22B Thinking90.6%$0.45$3.49
    5Claude 3.5 Sonnet90.4%$3.00$15.00
    6Claude 3.5 Sonnet90.4%$3.00$15.00
    7Kimi K2 090590.2%$0.60$2.50
    8GPT-4.190.2%$2.00$8.00

    Related LLM leaderboards

    LLM LeaderboardCodingMathAgenticReasoningWritingScienceFinanceLong ContextVisionOpen Source

    Individual benchmark pages

    Rank one eval at a time. All LLM benchmarks.

    HLESWE BenchHumanity's Last Exam (No Tools)Humanity's Last Exam (With Tools)Terminal-Bench 2.0Terminal-BenchTerminal-Bench 2.1Terminal-Bench 2Terminal-Bench 3.0ARC-AGI-1 Verified

    MMLU FAQ

    Who ranks #1 on the MMLU leaderboard?

    As of August 20, 2026, GPT-5 by OpenAI ranks #1 on MMLU at 92.5%. API pricing is $1.25/M input and $10.00/M output.

    What are the top models on MMLU?

    The current MMLU ranking as of August 20, 2026 is 1. GPT-5 at 92.5%; 2. o1 at 91.8%; 3. o1-preview at 90.8%.

    Which mmlu model is the cheapest?

    Llama 3.2 3B Instruct is the cheapest scored model on this mmlu leaderboard at $0.01/M input and $0.02/M output ($0.03 blended). GPT-5 still leads MMLU at 92.5%.

    Should I always pick the #1 MMLU model?

    Not automatically. GPT-5 leads MMLU, but a cheaper scored model can be the better production choice if the quality gap is small. Use the table to weigh MMLU against input/output price, context window, and related evals.

    How often is the MMLU leaderboard updated?

    Scores and API prices on this page are refreshed from published evals and provider rates. The snapshot is labeled August 20, 2026. Treat it as a current index, not a one-off blog post.

    What is MMLU?

    Massive Multitask Language Understanding — tests knowledge across 57 subjects including STEM, humanities, and social sciences. This page ranks models that have published a MMLU score, with live API token prices on the same row. Official methodology: MMLU (https://github.com/hendrycks/test).

    Where is the MMLU leaderboard?

    This page is the MMLU leaderboard. Models are sorted by MMLU, with input and output token prices on the same row so you can weigh score against cost. Official boards often omit price; that comparison is the point of this index.

    How is this MMLU ranking different from the official board?

    The official MMLU page owns the methodology. This page keeps the published MMLU score next to live API $/M so you can pick a production SKU, not only a trophy number. Source: https://github.com/hendrycks/test