TableVista: leaderboard
We introduce TableVista, a comprehensive benchmark for evaluating foundation models in multimodal table reasoning under visual and structural complexity. TableVista consists of 3,000 high-quality table reasoning problems, where each instance is expanded into 10 distinct visual variants through our multi-style rendering and transformation pipeline. This process encompasses diverse scenario styles,
Metric: Avg. (self-reported). Source: benchmarklist.com. Saturation forecast: Around June 2027. 29 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.4 | 72.1 | #79 |
| 2 | Llama 4 Maverick | 54.7 | #464 |
| 3 | Gemma 4 31B (IT) | 54.3 | #176 |
| 4 | Qwen 3.5 27B | 51.4 | #187 |
| 5 | Qwen 3.5 122B A10B | 51.1 | #173 |
| 6 | GPT-5.4 Mini | 49.4 | #212 |
| 7 | Llama 4 Scout | 49.4 | #656 |
| 8 | Qwen 2.5 VL 7B | 44.6 | #643 |
| 9 | Qwen 3.6 35B A3B | 43.4 | #226 |
| 10 | Gemma 3 27B (IT) | 42.8 | #523 |
| 11 | Gemma 4 26B A4B (IT) | 42.5 | #297 |
| 12 | Qwen 3.5 9B | 40.7 | #379 |
| 13 | Qwen 3.5 35B A3B | 40 | #256 |
| 14 | Molmo2-8B | 39.9 | #543 |
| 15 | Gemma 3 12B (IT) | 38.5 | #664 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=tablevista · How It Works · Data refreshed daily, snapshot 2026-10-20.