# MMMU leaderboard

Multimodal & Grounded benchmark

> Every figure below is reproduced as its upstream source published it: nothing is modelled, estimated, interpolated or converted. `Unavailable` means no source published the value — it is never a zero. Each value carries its evidence state and the date it was observed.

Page: https://modelscale.dev/leaderboards/benchmarks/multimodal-grounded-mmmu

Massive Multi-discipline Multimodal Understanding. Every model the catalog carries a published MMMU value for, ranked by that value.

## Benchmark record

| Field | Value |
| --- | --- |
| Category | Multimodal & Grounded |
| Measure | Image + text question answering |
| Tasks | Multimodal academic reasoning |
| Difficulty | Frontier multimodal |
| Published by | [MMMU: A Massive Multi-discipline Multimodal Understanding and Reasoning Benchmark for Expert AGI](https://arxiv.org/abs/2401.05508) |

A broad multimodal reasoning benchmark spanning charts, diagrams, tables, and academic visual question answering.

## MMMU ranking

32 models with a published MMMU value, ordered by that value, highest first. Models the source has not scored on this benchmark are not listed — they are unmeasured, not last.

| Rank | Model | Provider | Image + text question answering | Evidence |
| --- | --- | --- | --- | --- |
| 1 | [Gemini 3 Pro](https://modelscale.dev/models/gemini-3-pro) | Google | 87.2 | Observed 2026-11-21 · source benchlm:models |
| 2 | [GPT-5.2](https://modelscale.dev/models/gpt-5-2) | OpenAI | 86.7 | Observed 2026-11-21 · source benchlm:models |
| 3 | [Qwen3.6 Plus](https://modelscale.dev/models/qwen3-6-plus) | Alibaba | 86.0 | Observed 2026-11-21 · source benchlm:models |
| 4 | [Qwen3.5 397B](https://modelscale.dev/models/qwen3-5-397b) | Alibaba | 85.0 | Observed 2026-11-21 · source benchlm:models |
| 5 | [Kimi K2.5](https://modelscale.dev/models/kimi-k2-5) | Moonshot AI | 84.3 | Observed 2026-11-21 · source benchlm:models |
| 6 | [Qwen3.5-122B-A10B](https://modelscale.dev/models/qwen3-5-122b-a10b) | Alibaba | 83.9 | Observed 2026-11-21 · source benchlm:models |
| 7 | [Qwen3.6-27B](https://modelscale.dev/models/qwen3-6-27b) | Alibaba | 82.9 | Observed 2026-11-21 · source benchlm:models |
| 8 | [Qwen3.5-27B](https://modelscale.dev/models/qwen3-5-27b) | Alibaba | 82.3 | Observed 2026-11-21 · source benchlm:models |
| 9 | [Qwen3.6-35B-A3B](https://modelscale.dev/models/qwen3-6-35b-a3b) | Alibaba | 81.7 | Observed 2026-11-21 · source benchlm:models |
| 10 | [Qwen3.5-35B-A3B](https://modelscale.dev/models/qwen3-5-35b-a3b) | Alibaba | 81.4 | Observed 2026-11-21 · source benchlm:models |
| 11 | [Claude Opus 4.5](https://modelscale.dev/models/claude-opus-4-5) | Anthropic | 80.7 | Observed 2026-11-21 · source benchlm:models |
| 12 | [Qwen3-VL-235B-A22B Instruct](https://modelscale.dev/models/qwen3-vl-235b-a22b-instruct) | Alibaba | 78.7 | Observed 2026-11-21 · source benchlm:models |
| 13 | [Claude 4 Opus Thinking](https://modelscale.dev/models/claude-4-opus-thinking) | Anthropic | 76.5 | Observed 2026-11-21 · source benchlm:models |
| 14 | [GLM-4.6V](https://modelscale.dev/models/glm-4-6v) | Z.AI | 76.0 | Observed 2026-11-21 · source benchlm:models |
| 15 | [Command A+](https://modelscale.dev/models/command-a-plus) | Cohere | 75.1 | Observed 2026-11-21 · source benchlm:models |
| 16 | [Claude 3.7 Sonnet Thinking](https://modelscale.dev/models/claude-3-7-sonnet-thinking) | Anthropic | 75.0 | Observed 2026-11-21 · source benchlm:models |
| 17 | [Claude 4 Opus](https://modelscale.dev/models/claude-4-opus) | Anthropic | 73.7 | Observed 2026-11-21 · source benchlm:models |
| 18 | [Gemini 2.5 Flash-Lite](https://modelscale.dev/models/gemini-2-5-flash-lite) | Google | 72.9 | Observed 2026-11-21 · source benchlm:models |
| 19 | [Claude 3.7 Sonnet](https://modelscale.dev/models/claude-3-7-sonnet) | Anthropic | 71.8 | Observed 2026-11-21 · source benchlm:models |
| 20 | [GLM-4.6V-Flash](https://modelscale.dev/models/glm-4-6v-flash) | Z.AI | 71.1 | Observed 2026-11-21 · source benchlm:models |
| 21 | [Nemotron 3 Nano Omni 30B A3B](https://modelscale.dev/models/nemotron-3-nano-omni-30b-a3b) | NVIDIA | 70.8 | Observed 2026-11-21 · source benchlm:models |
| 22 | [Gemini 2.0 Flash](https://modelscale.dev/models/gemini-2-0-flash) | Google | 69.3 | Observed 2026-11-21 · source benchlm:models |
| 23 | [MiniMax-VL-01](https://modelscale.dev/models/minimax-vl-01) | MiniMax | 68.5 | Observed 2026-11-21 · source benchlm:models |
| 24 | [Grok 2](https://modelscale.dev/models/grok-2) | xAI | 66.1 | Observed 2026-11-21 · source benchlm:models |
| 25 | [Gemini 2.0 Flash-Lite](https://modelscale.dev/models/gemini-2-0-flash-lite) | Google | 65.1 | Observed 2026-11-21 · source benchlm:models |
| 26 | [Mistral Small 3.1](https://modelscale.dev/models/mistral-small-3-1) | Mistral | 64.0 | Observed 2026-11-21 · source benchlm:models |
| 27 | [Grok 2 mini](https://modelscale.dev/models/grok-2-mini) | xAI | 63.2 | Observed 2026-11-21 · source benchlm:models |
| 28 | [Gemini 1.5 Flash](https://modelscale.dev/models/gemini-1-5-flash) | Google | 58.3 | Observed 2026-11-21 · source benchlm:models |
| 29 | [LFM3-VL-2B](https://modelscale.dev/models/lfm3-vl-2b) | LiquidAI | 55.7 | Observed 2026-11-21 · source benchlm:models |
| 30 | [LFM2.5-VL-3B](https://modelscale.dev/models/lfm2-5-vl-3b) | LiquidAI | 48.4 | Observed 2026-11-21 · source benchlm:models |
| 31 | [ZAYA1-VL-8B](https://modelscale.dev/models/zaya1-vl-8b) | Zyphra | 46.0 | Observed 2026-11-21 · source benchlm:models |
| 32 | [LFM2.5-VL-450M](https://modelscale.dev/models/lfm2-5-vl-450m) | LiquidAI | 32.7 | Observed 2026-11-21 · source benchlm:models |

Rows are ordered by the value the source published, highest first. The source does not state whether a higher value is the better result, so this document does not either: for a benchmark that measures a rate of failure, read the table from the bottom.
