MMMU-Pro with Python (MMMU-Pro w/ Python)
We show this table for reference; we do not rank on it.
Tool-augmented MMMU-Pro variant that allows Python assistance during multimodal reasoning.
Benchmark score on MMMU-Pro w/ Python — November 13, 2026
We compile the MMMU-Pro w/ Python rows from provider self-reports. Kimi K3.1 Pro leads the table at 90%, followed by GPT-5.6 Sol (84.6%) and Kimi K3 (83.4%). We do not use these results to rank models overall.
Kimi K3.1 Pro
Moonshot AI
Reported by Moonshot AIGPT-5.6 Sol
OpenAI
Reported by OpenAIKimi K3
Moonshot AI
Reported by Moonshot AI10 modelsMultimodal & GroundedCurrentDisplay onlyUpdated November 13, 2026
Benchmark score table (10 models)
All 10 results on this page were reported by the labs that built the models.
| Rank | Model / configuration | Score | Who ran it | Parameters (B) | Open / closed |
|---|---|---|---|---|---|
| 1 | Kimi K3.1 ProMoonshot AIReported by Moonshot AI | 90% | Reported by Moonshot AI | Not reported | Open |
| 2 | GPT-5.6 SolOpenAIReported by OpenAI | 84.6% | Reported by OpenAI | Not reported | Closed |
| 3 | Kimi K3Moonshot AIReported by Moonshot AI | 83.4% | Reported by Moonshot AI | Not reported | Open |
| 4 | GPT-5.5OpenAIReported by OpenAI | 83.2% | Reported by OpenAI | Not reported | Closed |
| 5 | GPT-5.4OpenAIReported by OpenAI | 82.1% | Reported by OpenAI | Not reported | Closed |
| 6 | GPT-5.6 TerraOpenAIReported by OpenAI | 82% | Reported by OpenAI | Not reported | Closed |
| 7 | Kimi K2.6Moonshot AIReported by Moonshot AI | 80.1% | Reported by Moonshot AI | Not reported | Open |
| 8 | GPT-5.6 LunaOpenAIReported by OpenAI | 79.5% | Reported by OpenAI | Not reported | Closed |
| 9 | GPT-5.4 miniOpenAIReported by OpenAI | 78% | Reported by OpenAI | Not reported | Closed |
| 10 | GPT-5.4 nanoOpenAIReported by OpenAI | 69.5% | Reported by OpenAI | Not reported | Closed |
- Lab-reported
- Benchmark-run
- Independent
Among the reported MMMU-Pro w/ Python rows, Kimi K3.1 Pro is first at 90%. The third row is 6.6 points behind. The broader top-10 range is 20.5 points, so the table still separates the published systems.
10 models have been evaluated on MMMU-Pro w/ Python. The benchmark falls in the Multimodal & Grounded category. MMMU-Pro w/ Python is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.
About MMMU-Pro w/ Python
Year
2026
Tasks
Multimodal academic reasoning
Format
Image + text question answering with Python
Difficulty
Frontier multimodal
Useful for measuring multimodal reasoning when the model can combine visual understanding with computation.
Freshness and provenance
Version
MMMU-Pro w/ Python 2026
Refresh cadence
Quarterly
Staleness state
Current
Question availability
Public benchmark set
BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.
Questions
What does MMMU-Pro w/ Python measure?
Tool-augmented MMMU-Pro variant that allows Python assistance during multimodal reasoning.
Which model scores highest on MMMU-Pro w/ Python?
Kimi K3.1 Pro by Moonshot AI currently leads with a score of 90% on MMMU-Pro w/ Python.
How many models are evaluated on MMMU-Pro w/ Python?
10 AI models have published results on MMMU-Pro w/ Python in the BenchLM catalog.
Compare top models on MMMU-Pro w/ Python
Know when it’s worth switching models
The model to choose, the cheaper alternative, and the release we would wait on.
Read a sample issueJoin 7,000+ subscribers.
One email each week. Unsubscribe anytime.