Multimodal & Grounded benchmark
MMMU-Pro leaderboard
Massive Multi-discipline Multimodal Understanding Pro. Every model the catalog carries a published MMMU-Pro value for, ranked by that value.
A harder multimodal benchmark for frontier models that combines text with images, diagrams, charts, and academic visual reasoning tasks.
MMMU-Pro ranking
41 models with a published MMMU-Pro 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 |
|---|---|---|---|
| 1 |  Gemini 3.1 Pro | 83.9 | |
| 2 | Gemini 3.5 Flash | 83.6 | |
| 3 |  GPT-5.6 Sol | OpenAI | 83 |
| 4 |  Qwen3.8 Max | Alibaba | 82.3 |
| 5 |  Kimi K3 | Moonshot AI | 81.6 |
| 5 |  Seed 2.1 Pro | ByteDance | 81.6 |
| 7 | GPT-5.4 | OpenAI | 81.2 |
| 7 | GPT-5.5 | OpenAI | 81.2 |
| 9 | Gemini 3 Pro | 81 | |
| 10 | GPT-5.6 Terra | OpenAI | 80.7 |
| 11 |  Muse Spark | Meta | 80.4 |
| 12 | Seed 2.1 Turbo | ByteDance | 80.1 |
| 13 | GPT-5.2 | OpenAI | 79.5 |
| 14 | Kimi K2.6 | Moonshot AI | 79.4 |
| 15 | DSdots3-note Preview | Dots Studio | 79.1 |
| 16 | Qwen3.5 397B | Alibaba | 79 |
| 16 | Qwen3.7 Plus | Alibaba | 79 |
| 18 | Qwen3.6 Plus | Alibaba | 78.8 |
| 19 | Kimi K2.5 | Moonshot AI | 78.5 |
| 19 | Kimi K2.5 (Reasoning) | Moonshot AI | 78.5 |
| 21 | GPT-5.6 Luna | OpenAI | 78.4 |
| 22 |  Grok 4.3 | xAI | 78.1 |
| 22 |  MiniMax M3 | MiniMax | 78.1 |
| 24 | UNPareto 26.9 | Unbiased | 78 |
| 25 |  MiMo-V2.5 | Xiaomi | 77.9 |
| 26 |  Claude Opus 4.6 | Anthropic | 77.3 |
| 27 | Gemma 4 31B | 76.9 | |
| 28 | GPT-5.4 mini | OpenAI | 76.6 |
| 29 | Qwen3.6-27B | Alibaba | 75.8 |
| 30 | Qwen3.6-35B-A3B | Alibaba | 75.3 |
| 31 | Grok 4.20 | xAI | 75.2 |
| 32 | TMInkling-Small | Thinking Machines Lab | 74 |
| 32 | Muse Glimmer 30B | Meta | 74 |
| 34 | Gemma 4 26B A4B | 73.8 | |
| 35 | TMInkling | Thinking Machines Lab | 73.5 |
| 36 | INInterfaze Beta | Interfaze | 71.1 |
| 37 | Claude Opus 4.5 | Anthropic | 70.6 |
| 38 | Gemma 4 12B | 69.1 | |
| 39 | GPT-5.4 nano | OpenAI | 66.1 |
| 40 |  Command A+ | Cohere | 63 |
| 41 |  LFM2.5-VL-3B | LiquidAI | 30.5 |
Evidence key: Observed
Rows are ordered by the value MMMU-Pro: A More Robust Multi-discipline Multimodal Understanding Benchmark published, highest first. The source does not state whether a higher value is the better result, so this page does not either: for a benchmark that measures a rate of failure, read the table from the bottom.