MMTU - Column Transform: leaderboard
The "Column Transform" subtask of MMTU. Massive Multi-Task Table Understanding benchmark with real-world table reasoning tasks including joins, transformations, cleaning, and question answering.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Around February 2027. 26 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5 | 52.05 | #98 |
| 2 | O3 | 51.99 | #124 |
| 3 | GPT-5 Mini | 51.34 | #186 |
| 4 | Gemini 2.5 Pro | 48.58 | #152 |
| 5 | Grok 3 Mini | 46.48 | #281 |
| 6 | GPT-OSS-120B | 46.34 | #341 |
| 7 | DeepSeek R1 0528 | 45.06 | #222 |
| 8 | Gemini 2.5 Flash | 43.84 | #245 |
| 9 | Qwen 3 235B A22B 2507 (Thinking) | 43.72 | #260 (Qwen 3 235B A22B 2507) |
| 10 | GPT-5 Chat | 41.62 | #241 |
| 11 | Qwen 3 32B | 39.75 | #443 |
| 12 | Qwen 3 235B A22B 2507 Instruct | 37.76 | #300 |
| 13 | GPT-OSS-20B | 37.58 | #518 |
| 14 | GPT-4o (2024-11-20) | 36.89 | #381 |
| 15 | Mistral Large 2 (Nov) Instruct (2411) | 36.72 | #465 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=mmtu-column-transform · How It Works · Data refreshed daily, snapshot 2026-10-20.