Multimodal & Grounded benchmark
MMMU-Pro w/ Python leaderboard
MMMU-Pro with Python. Every model the catalog carries a published MMMU-Pro w/ Python value for, ranked by that value.
CategoryMultimodal & Grounded
MeasureImage + text question answering with Python
TasksMultimodal academic reasoning
DifficultyFrontier multimodal
Published byIntroducing GPT-5.4 mini and nano
Tool-augmented MMMU-Pro variant that allows Python assistance during multimodal reasoning.
MMMU-Pro w/ Python ranking
9 models with a published MMMU-Pro w/ Python 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 with Python |
|---|---|---|---|
| 1 |  GPT-5.6 Sol | OpenAI | 84.6 |
| 2 |  Kimi K3 | Moonshot AI | 83.4 |
| 3 | GPT-5.5 | OpenAI | 83.2 |
| 4 | GPT-5.4 | OpenAI | 82.1 |
| 5 | GPT-5.6 Terra | OpenAI | 82 |
| 6 | Kimi K2.6 | Moonshot AI | 80.1 |
| 7 | GPT-5.6 Luna | OpenAI | 79.5 |
| 8 | GPT-5.4 mini | OpenAI | 78 |
| 9 | GPT-5.4 nano | OpenAI | 69.5 |
Evidence key: Observed
Rows are ordered by the value Introducing GPT-5.4 mini and nano 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.