# Vals MMMU (MMMU)

> Multimodal Multi-task Benchmark

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

- Category: [Multimodal & Grounded](/multimodal-grounded)
- Last updated: September 1, 2026

## About MMMU

- Year: 2026
- Tasks: Multimodal academic task suite
- Format: Accuracy score
- Difficulty: Multimodal college-level reasoning
- Paper: [MMMU](https://www.vals.ai/benchmarks/mmmu)

BenchLM mirrors the public Vals AI MMMU leaderboard as display-only external evidence. The captured snapshot preserves overall scores, task-level scores where Vals publishes them, uncertainty, latency, and cost-per-test metadata. It is excluded from BenchLM weighted rankings.

MMMU is currently displayed on BenchLM for reference, but it is excluded from the weighted scoring formula.

## Leaderboard (93 models)

| Rank | Model | Creator | Score |
|------|-------|---------|-------|
| 1 | [Claude Fable 5.1](/models/claude-fable-5-1) | Anthropic | 90.64% |
| 2 | [Claude Opus 5](/models/claude-opus-5) | Anthropic | 89.88% |
| 3 | [Claude Fable 5](/models/claude-fable) | Anthropic | 89.31% |
| 4 | [Gemini 3.8 Flash](/models/gemini-3-8-flash) | Google | 89.08% |
| 5 | [Gemini 3.7 Flash](/models/gemini-3-7-flash) | Google | 88.96% |
| 6 | [GPT-5.6 Sol](/models/gpt-5-6-sol) | OpenAI | 88.84% |
| 7 | [Gemini 3.6 Flash](/models/gemini-3-6-flash) | Google | 88.38% |
| 8 | [Gemini 3.5 Flash](/models/gemini-3-5-flash) | Google | 88.27% |
| 9 | [GPT-5.5](/models/gpt-5-5) | OpenAI | 88.27% |
| 10 | [Gemini 3.1 Pro Preview](https://www.vals.ai/models/google_gemini-3.1-pro-preview) | Google | 88.21% |
| 11 | [Kimi K3](/models/kimi-k3) | Moonshot AI | 88.15% |
| 12 | [Qwen3.8 Max](/models/qwen3-8-max) | Alibaba | 88.03% |
| 13 | [Gemini 3 Flash Preview](https://www.vals.ai/models/google_gemini-3-flash-preview) | Google | 87.63% |
| 14 | [Gemini 3 Pro Preview](https://www.vals.ai/models/google_gemini-3-pro-preview) | Google | 87.51% |
| 15 | [GPT-5.4](/models/gpt-5-4) | OpenAI | 87.51% |
| 16 | [Muse Spark](/models/muse-spark) | Meta | 87.40% |
| 17 | [GPT-5.2](/models/gpt-5-2) | OpenAI | 86.67% |
| 18 | [Muse Spark 1.1](/models/muse-spark-1-1) | Meta | 86.59% |
| 19 | [Claude Opus 4.8](/models/claude-opus-4-8) | Anthropic | 86.59% |
| 20 | [GPT-5.6 Terra](/models/gpt-5-6-terra) | OpenAI | 86.47% |
| 21 | [Kimi K2.6](/models/kimi-2-6) | Moonshot AI | 86.30% |
| 22 | [Muse Spark 1.2](/models/muse-spark-1-2) | Meta | 86.13% |
| 23 | [GLM-5.3-Flash](/models/glm-5-3-flash) | Z.AI | 86.01% |
| 24 | [Claude Opus 4.7](/models/claude-opus-4-7) | Anthropic | 85.55% |
| 25 | [GPT-5.6 Luna](/models/gpt-5-6-luna) | OpenAI | 85.03% |
| 26 | [Kimi K2.5 Thinking](https://www.vals.ai/models/kimi_kimi-k2.5-thinking) | Moonshot AI | 84.33% |
| 27 | [Qwen3.6 Plus](/models/qwen3-6-plus) | Alibaba | 84.16% |
| 28 | [Qwen3.8-27B](/models/qwen3-8-27b) | Alibaba | 83.93% |
| 29 | [Claude Opus 4.6 (Adaptive)](/models/claude-opus-4-6-thinking) | Anthropic | 83.87% |
| 30 | [Gemini 3.5 Flash-Lite](/models/gemini-3-5-flash-lite) | Google | 83.58% |
| 31 | [Claude Sonnet 4.6](/models/claude-sonnet-4-6) | Anthropic | 83.58% |
| 32 | [Grok 4.20 0309 Reasoning](https://www.vals.ai/models/grok_grok-4.20-0309-reasoning) | xAI | 83.47% |
| 33 | [GPT-5.1](/models/gpt-5-1) | OpenAI | 83.18% |
| 34 | [Grok 4.3](/models/grok-4-3) | xAI | 83.06% |
| 35 | [Claude Sonnet 5](/models/claude-sonnet-5) | Anthropic | 83.01% |
| 36 | [Claude Opus 4.5 Thinking](/models/claude-opus-4-5-thinking) | Anthropic | 82.95% |
| 37 | [Gemini 3.1 Flash Lite Preview](https://www.vals.ai/models/google_gemini-3.1-flash-lite-preview) | Google | 82.49% |
| 38 | [Qwen3.5 Flash](/models/qwen3-5-flash) | Alibaba | 81.91% |
| 39 | [GPT-5](https://www.vals.ai/models/openai_gpt-5-2025-08-07) | OpenAI | 81.50% |
| 40 | [Gemini 2.5 Pro Exp 03 25](https://www.vals.ai/models/google_gemini-2.5-pro-exp-03-25) | Google | 81.34% |
| 41 | [MiniMax M3](/models/minimax-m3) | MiniMax | 81.16% |
| 42 | [Claude Opus 4.5](/models/claude-opus-4-5) | Anthropic | 81.10% |
| 43 | [Gemini 2.5 Flash Preview 09 2025 Thinking](https://www.vals.ai/models/google_gemini-2.5-flash-preview-09-2025-thinking) | Google | 80.75% |
| 44 | [o3](/models/o3) | OpenAI | 80.42% |
| 45 | [MiMo-V2.5](/models/mimo-v2-5) | Xiaomi | 80.00% |
| 46 | [O4 Mini](https://www.vals.ai/models/openai_o4-mini-2025-04-16) | OpenAI | 79.67% |
| 47 | [Gemini 2.5 Flash Preview 09 2025](https://www.vals.ai/models/google_gemini-2.5-flash-preview-09-2025) | Google | 79.48% |
| 48 | [Claude Sonnet 4.5 Thinking](/models/claude-sonnet-4-5-thinking) | Anthropic | 79.31% |
| 49 | [GPT-5.4 mini](/models/gpt-5-4-mini) | OpenAI | 79.25% |
| 50 | [GPT-5 mini](/models/gpt-5-mini) | OpenAI | 78.91% |
| 51 | [Inkling](/models/inkling) | Thinking Machines Lab | 78.50% |
| 52 | [Claude Opus 4.1 20250805 Thinking](https://www.vals.ai/models/anthropic_claude-opus-4-1-20250805-thinking) | Anthropic | 77.51% |
| 53 | [o1](/models/o1) | OpenAI | 77.41% |
| 54 | [Grok 4 0709](https://www.vals.ai/models/grok_grok-4-0709) | xAI | 76.27% |
| 55 | [Gemini 2.5 Flash Lite Preview 09 2025 Thinking](https://www.vals.ai/models/google_gemini-2.5-flash-lite-preview-09-2025-thinking) | Google | 75.43% |
| 56 | [Claude 3.7 Sonnet 20250219 Thinking](https://www.vals.ai/models/anthropic_claude-3-7-sonnet-20250219-thinking) | Anthropic | 75.10% |
| 57 | [Claude Sonnet 4 20250514 Thinking](https://www.vals.ai/models/anthropic_claude-sonnet-4-20250514-thinking) | Anthropic | 74.93% |
| 58 | [Claude Opus 4.1](https://www.vals.ai/models/anthropic_claude-opus-4-1-20250805) | Anthropic | 73.72% |
| 59 | [GPT-5.4 nano](/models/gpt-5-4-nano) | OpenAI | 73.58% |
| 60 | [Claude Opus 4](https://www.vals.ai/models/anthropic_claude-opus-4-20250514) | Anthropic | 73.31% |
| 61 | [Grok 4 Fast (Reasoning)](/models/grok-4-fast-reasoning) | xAI | 72.78% |
| 62 | [Grok 4.1 Fast (Reasoning)](/models/grok-4-1-fast-reasoning) | xAI | 72.66% |
| 63 | [Gemini 2.5 Flash Lite Preview 09 2025](https://www.vals.ai/models/google_gemini-2.5-flash-lite-preview-09-2025) | Google | 72.54% |
| 64 | [Claude Sonnet 4](https://www.vals.ai/models/anthropic_claude-sonnet-4-20250514) | Anthropic | 72.39% |
| 65 | [GPT-4.1](/models/gpt-4-1) | OpenAI | 72.39% |
| 66 | [Gemini 2.5 Flash Preview 04 17 Thinking](https://www.vals.ai/models/google_gemini-2.5-flash-preview-04-17-thinking) | Google | 71.92% |
| 67 | [Llama4 Maverick Instruct Basic](https://www.vals.ai/models/fireworks_llama4-maverick-instruct-basic) | Fireworks | 71.69% |
| 68 | [Claude 3.7 Sonnet](https://www.vals.ai/models/anthropic_claude-3-7-sonnet-20250219) | Anthropic | 71.52% |
| 69 | [GPT-5 nano](/models/gpt-5-nano) | OpenAI | 70.94% |
| 70 | [GPT-4.1 mini](/models/gpt-4-1-mini) | OpenAI | 70.54% |
| 71 | [Gemini 2.0 Flash 001](https://www.vals.ai/models/google_gemini-2.0-flash-001) | Google | 69.79% |
| 72 | [Claude 3.5 Sonnet](/models/claude-3-5-sonnet) | Anthropic | 68.80% |
| 73 | [Gemini 2.5 Flash Preview 04 17](https://www.vals.ai/models/google_gemini-2.5-flash-preview-04-17) | Google | 68.00% |
| 74 | [Mistral Large 2512](https://www.vals.ai/models/mistralai_mistral-large-2512) | Mistral | 66.19% |
| 75 | [Gemini 1.5 Pro 002](https://www.vals.ai/models/google_gemini-1.5-pro-002) | Google | 65.51% |
| 76 | [Magistral Small 2509](https://www.vals.ai/models/mistralai_magistral-small-2509) | Mistral | 65.20% |
| 77 | [Magistral Medium 2509](https://www.vals.ai/models/mistralai_magistral-medium-2509) | Mistral | 64.57% |
| 78 | [GPT-4o](/models/gpt-4o) | OpenAI | 64.01% |
| 79 | [Grok 4.1 Fast Non Reasoning](https://www.vals.ai/models/grok_grok-4-1-fast-non-reasoning) | xAI | 63.70% |
| 80 | [Grok 4 Fast Non Reasoning](https://www.vals.ai/models/grok_grok-4-fast-non-reasoning) | xAI | 63.41% |
| 81 | [Mistral Medium 2505](https://www.vals.ai/models/mistralai_mistral-medium-2505) | Mistral | 62.97% |
| 82 | [GPT-4o](/models/gpt-4o) | OpenAI | 62.16% |
| 83 | [Grok 4.5](/models/grok-4-5) | xAI | 61.79% |
| 84 | [Mistral Small 2503](https://www.vals.ai/models/mistralai_mistral-small-2503) | Mistral | 60.08% |
| 85 | [Meta Llama Llama 4 Scout 17B 16E Instruct](https://www.vals.ai/models/together_meta-llama/Llama-4-Scout-17B-16E-Instruct) | Together AI | 58.75% |
| 86 | [Grok 2 Vision 1212](https://www.vals.ai/models/grok_grok-2-vision-1212) | xAI | 57.25% |
| 87 | [Gemini 1.5 Flash 002](https://www.vals.ai/models/google_gemini-1.5-flash-002) | Google | 57.19% |
| 88 | [GPT-4o mini](/models/gpt-4o-mini) | OpenAI | 56.56% |
| 89 | [GPT-4.1 nano](/models/gpt-4-1-nano) | OpenAI | 55.05% |
| 90 | [Meta Llama Llama 3.2 90B Vision Instruct Turbo](https://www.vals.ai/models/together_meta-llama/Llama-3.2-90B-Vision-Instruct-Turbo) | Together AI | 48.06% |
| 91 | [Claude Haiku 4.5 Thinking](/models/claude-haiku-4-5-thinking) | Anthropic | 46.07% |
| 92 | [Meta Llama Llama 3.2 11B Vision Instruct Turbo](https://www.vals.ai/models/together_meta-llama/Llama-3.2-11B-Vision-Instruct-Turbo) | Together AI | 38.82% |
| 93 | [Qwen3.5 Plus Thinking](https://www.vals.ai/models/alibaba_qwen3.5-plus-thinking) | Alibaba | 22.77% |

## FAQ

### What does MMMU measure?

Multimodal Multi-task Benchmark

### Which model leads the published MMMU snapshot?

Claude Fable 5.1 currently leads the published MMMU snapshot with a score of 90.64%.

### How many models are evaluated on MMMU?

The September 1, 2026 contains 93 AI models.

### Does MMMU affect BenchLM's overall score?

Not directly. MMMU is still displayed on BenchLM for reference, but it is excluded from the weighted scoring formula.