# Artificial Analysis MMMU-Pro (AA-MMMU-Pro)

> An Artificial Analysis MMMU-Pro score.

Canonical page: https://benchlm.ai/benchmarks/aammmupro

- Category: [Multimodal & Grounded](/multimodal-grounded)
- Last updated: September 27, 2026

## About AA-MMMU-Pro

- Year: 2026
- Tasks: Multimodal academic reasoning
- Format: Image + text question answering
- Difficulty: Frontier multimodal
- Paper: [Artificial Analysis MMMU-Pro Benchmark Leaderboard](https://artificialanalysis.ai/evaluations/mmmu-pro)

BenchLM stores the Artificial Analysis MMMU-Pro result separately from the core MMMU-Pro lane because Artificial Analysis controls the evaluation configuration.

The Multimodal & Grounded leaderboard ranks models by a weighted category score, and AA-MMMU-Pro contributes 15% of it. It does not enter the overall BenchAlign v5.7 ranking.

## Leaderboard (101 models)

| Rank | Model | Creator | Score |
|------|-------|---------|-------|
| 1 | [Claude Opus 5.5](/models/claude-opus-5-5) | Anthropic | 87.7% |
| 2 | [GPT-6 Astra](/models/gpt-6-astra) | OpenAI | 86.9% |
| 3 | [Gemini 3.8 Flash](/models/gemini-3-8-flash) | Google | 85.6% |
| 4 | [Gemini 3.7 Flash](/models/gemini-3-7-flash) | Google | 85.5% |
| 5 | [Claude Opus 5](/models/claude-opus-5) | Anthropic | 84.7% |
| 6 | [Gemini 3.5 Flash](/models/gemini-3-5-flash) | Google | 84.3% |
| 7 | [GPT-5.6 Sol](/models/gpt-5-6-sol) | OpenAI | 83.4% |
| 8 | [GPT-6 Sol](/models/gpt-6-sol) | OpenAI | 83.3% |
| 9 | [Gemini 3.6 Flash](/models/gemini-3-6-flash) | Google | 83.2% |
| 10 | [Qwen3.8 Max Preview](/models/qwen3-8-max-preview) | Alibaba | 82.8% |
| 11 | [Gemini 3.1 Pro](/models/gemini-3-1-pro) | Google | 82.4% |
| 12 | [GPT-5.6 Terra](/models/gpt-5-6-terra) | OpenAI | 80.7% |
| 13 | [Kimi K3](/models/kimi-k3) | Moonshot AI | 80.5% |
| 14 | [Muse Spark](/models/muse-spark) | Meta | 80.5% |
| 15 | [Qwen3.7 Plus](/models/qwen3-7-plus) | Alibaba | 80.5% |
| 16 | [Grok 4.5](/models/grok-4-5) | xAI | 80.4% |
| 17 | [Gemini 3 Pro](/models/gemini-3-pro) | Google | 80.2% |
| 18 | [GPT-5.5](/models/gpt-5-5) | OpenAI | 79.9% |
| 19 | [Qwen3.8-Flash-Next](/models/qwen3-8-flash-next) | Alibaba | 79.8% |
| 20 | [Kimi K2.6](/models/kimi-2-6) | Moonshot AI | 79.4% |
| 21 | [Apodex 1.1](/models/apodex-1-1) | Apodex | 79.2% |
| 22 | [Apodex 1.1 Mini](/models/apodex-1-1-mini) | Apodex | 79.2% |
| 23 | [Gemini 3.5 Flash-Lite](/models/gemini-3-5-flash-lite) | Google | 79.0% |
| 24 | [Ling 3.0 Flash VL](/models/ling-3-0-flash-vl) | InclusionAI | 79.0% |
| 25 | [Claude Opus 4.7 (Adaptive)](/models/claude-opus-4-7-adaptive) | Anthropic | 78.8% |
| 26 | [GPT-5.6 Luna](/models/gpt-5-6-luna) | OpenAI | 78.6% |
| 27 | [Gemini 3 Flash](/models/gemini-3-flash) | Google | 78.6% |
| 28 | [MiniMax M3](/models/minimax-m3) | MiniMax | 78.6% |
| 29 | [GPT-5.3 Codex](/models/gpt-5-3-codex) | OpenAI | 78.5% |
| 30 | [GPT-5.4](/models/gpt-5-4) | OpenAI | 78.4% |
| 31 | [Grok 4.3](/models/grok-4-3) | xAI | 78.1% |
| 32 | [Qwen3.6 Plus](/models/qwen3-6-plus) | Alibaba | 78.0% |
| 33 | [Claude Sonnet 5](/models/claude-sonnet-5) | Anthropic | 77.3% |
| 34 | [DeepSeek V4.1 Flash](/models/deepseek-v4-1-flash) | DeepSeek | 77.0% |
| 35 | [Claude Opus 4.7](/models/claude-opus-4-7) | Anthropic | 76.4% |
| 36 | [Step 5 Preview](/models/step-5-preview) | StepFun | 76.4% |
| 37 | [Qwen3.8-27B](/models/qwen3-8-27b) | Alibaba | 76.3% |
| 38 | [GPT-5.2-Codex](/models/gpt-5-2-codex) | OpenAI | 76.3% |
| 39 | [GPT-6 Luna](/models/gpt-6-luna) | OpenAI | 75.5% |
| 40 | [GPT-5.1](/models/gpt-5-1) | OpenAI | 75.5% |
| 41 | [Kimi K2.5](/models/kimi-k2-5) | Moonshot AI | 75.4% |
| 42 | [Kimi K2.5 (Reasoning)](/models/kimi-k2-5-reasoning) | Moonshot AI | 75.4% |
| 43 | [Claude Opus 4.6 (Adaptive)](/models/claude-opus-4-6-thinking) | Anthropic | 75.4% |
| 44 | [Step 3.7 Flash](/models/step-3-7-flash) | StepFun | 75.3% |
| 45 | [Qwen3.5-27B](/models/qwen3-5-27b) | Alibaba | 75.0% |
| 46 | [Qwen3.6-35B-A3B](/models/qwen3-6-35b-a3b) | Alibaba | 75.0% |
| 47 | [Qwen3.5-122B-A10B](/models/qwen3-5-122b-a10b) | Alibaba | 75.0% |
| 48 | [Gemini 2.5 Pro](/models/gemini-2-5-pro) | Google | 74.9% |
| 49 | [Qwen3.6-27B](/models/qwen3-6-27b) | Alibaba | 74.6% |
| 50 | [GPT-5 (medium)](/models/gpt-5-medium) | OpenAI | 74.3% |
| 51 | [Muse Glimmer 30B](/models/muse-glimmer-30b) | Meta | 74.3% |
| 52 | [GPT-5 (high)](/models/gpt-5-high) | OpenAI | 74.2% |
| 53 | [Inkling-Small](/models/inkling-small) | Thinking Machines Lab | 74.0% |
| 54 | [Claude Opus 4.5 Thinking](/models/claude-opus-4-5-thinking) | Anthropic | 74.0% |
| 55 | [Inkling](/models/inkling) | Thinking Machines Lab | 73.5% |
| 56 | [Gemma 4 31B](/models/gemma-4-31b) | Google | 73.4% |
| 57 | [GPT-5.4 mini](/models/gpt-5-4-mini) | OpenAI | 73.3% |
| 58 | [GLM-5V-Turbo](/models/glm-5v-turbo) | Z.AI | 72.8% |
| 59 | [Qwen3.5-35B-A3B](/models/qwen3-5-35b-a3b) | Alibaba | 72.7% |
| 60 | [Claude Opus 4.6](/models/claude-opus-4-6) | Anthropic | 72.5% |
| 61 | [GPT-5.1-Codex](/models/gpt-5-1-codex) | OpenAI | 72.5% |
| 62 | [GPT-5.1-Codex-Max](/models/gpt-5-1-codex-max) | OpenAI | 72.5% |
| 63 | [Claude Opus 4.5](/models/claude-opus-4-5) | Anthropic | 71.2% |
| 64 | [Claude Sonnet 4.6](/models/claude-sonnet-4-6) | Anthropic | 70.6% |
| 65 | [o3](/models/o3) | OpenAI | 70.1% |
| 66 | [MiMo-V2-Omni](/models/mimo-v2-omni) | Xiaomi | 69.9% |
| 67 | [Gemma 4 12B](/models/gemma-4-12b) | Google | 69.7% |
| 68 | [Gemma 4 26B A4B](/models/gemma-4-26b-a4b) | Google | 69.2% |
| 69 | [Grok 4](/models/grok-4) | xAI | 68.8% |
| 70 | [Claude 4.1 Opus Thinking](/models/claude-4-1-opus-thinking) | Anthropic | 67.9% |
| 71 | [Gemini 2.5 Flash](/models/gemini-2-5-flash) | Google | 65.5% |
| 72 | [GPT-5.4 nano](/models/gpt-5-4-nano) | OpenAI | 65.4% |
| 73 | [Mistral Medium 3.5 128B](/models/mistral-medium-3-5-128b) | Mistral | 64.9% |
| 74 | [Grok 4.1 Fast (Reasoning)](/models/grok-4-1-fast-reasoning) | xAI | 63.3% |
| 75 | [Command A+](/models/command-a-plus) | Cohere | 63.2% |
| 76 | [Claude 4 Sonnet](/models/claude-4-sonnet) | Anthropic | 62.4% |
| 77 | [Llama 4 Maverick](/models/llama-4-maverick) | Meta | 62.1% |
| 78 | [Grok 4 Fast (Reasoning)](/models/grok-4-fast-reasoning) | xAI | 61.8% |
| 79 | [GPT-4.1](/models/gpt-4-1) | OpenAI | 61.2% |
| 80 | [Qwen3-Omni-30B-A3B-Thinking](/models/qwen3-omni-30b-a3b-thinking) | Alibaba | 60.2% |
| 81 | [GPT-4.1 mini](/models/gpt-4-1-mini) | OpenAI | 58.7% |
| 82 | [Mistral Small 4](/models/mistral-small-4) | Mistral | 56.8% |
| 83 | [Mistral Small 4 (Reasoning)](/models/mistral-small-4-reasoning) | Mistral | 56.8% |
| 84 | [Mistral Large 3](/models/mistral-large-3) | Mistral | 55.7% |
| 85 | [Qwen3-Omni-30B-A3B-Instruct](/models/qwen3-omni-30b-a3b-instruct) | Alibaba | 55.5% |
| 86 | [Gemini 1.5 Pro](/models/gemini-1-5-pro) | Google | 55.0% |
| 87 | [Nemotron 3 Nano Omni 30B A3B](/models/nemotron-3-nano-omni-30b-a3b) | NVIDIA | 53.2% |
| 88 | [Mistral Medium 3](/models/mistral-medium-3) | Mistral | 53.0% |
| 89 | [Llama 4 Scout](/models/llama-4-scout) | Meta | 52.9% |
| 90 | [Qwen3.5 397B](/models/qwen3-5-397b) | Alibaba | 52.7% |
| 91 | [Qwen3.5 397B (Reasoning)](/models/qwen3-5-397b-reasoning) | Alibaba | 52.7% |
| 92 | [Gemma 4 E4B](/models/gemma-4-e4b) | Google | 51.4% |
| 93 | [Grok 4.1 Fast](/models/grok-4-1-fast) | xAI | 48.4% |
| 94 | [Gemma 3 27B](/models/gemma-3-27b) | Google | 48.0% |
| 95 | [Gemma 4 E2B](/models/gemma-4-e2b) | Google | 44.6% |
| 96 | [Nova Pro](/models/nova-pro) | Amazon | 44.3% |
| 97 | [GPT-4o mini](/models/gpt-4o-mini) | OpenAI | 41.5% |
| 98 | [GPT-4.1 nano](/models/gpt-4-1-nano) | OpenAI | 40.1% |
| 99 | [Claude 3 Haiku](/models/claude-3-haiku) | Anthropic | 30.8% |
| 100 | [LFM2.5-VL-1.6B-Extract](/models/lfm2-5-vl-1-6b-extract) | LiquidAI | 26.5% |
| 101 | [Phi-4 Multimodal Instruct](/models/phi-4-multimodal-instruct) | Microsoft | 14.5% |

## FAQ

### What does AA-MMMU-Pro measure?

An Artificial Analysis MMMU-Pro score.

### Which model scores highest on AA-MMMU-Pro?

Claude Opus 5.5 by Anthropic currently leads with a score of 87.7% on AA-MMMU-Pro.

### How many models are evaluated on AA-MMMU-Pro?

101 AI models have been evaluated on AA-MMMU-Pro on BenchLM.

### Does AA-MMMU-Pro affect BenchLM's overall score?

Not the overall score. The Multimodal & Grounded leaderboard ranks models by a weighted category score, and AA-MMMU-Pro contributes 15% of it. It does not enter the overall BenchAlign v5.7 ranking.

## Compare Top Models on AA-MMMU-Pro

- [Claude Opus 5.5 vs GPT-6 Astra](/compare/claude-opus-5-5-vs-gpt-6-astra)
- [GPT-6 Astra vs Gemini 3.8 Flash](/compare/gemini-3-8-flash-vs-gpt-6-astra)
- [Gemini 3.8 Flash vs Gemini 3.7 Flash](/compare/gemini-3-7-flash-vs-gemini-3-8-flash)
- [Gemini 3.7 Flash vs Claude Opus 5](/compare/claude-opus-5-vs-gemini-3-7-flash)