TableVision - Compositional Arithmetic: leaderboard

TableVision (2026) category of questions over images of hierarchical tables: L3 analysis, compositional arithmetic over many cells (107 questions). Exact-match accuracy (%), zero-shot; higher is better.

Metric: Exact-match accuracy (%) on L3 analysis, compositional arithmetic over many cells (107 questions), zero-shot, the model answers directly from the table image; TableVision test set; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 10 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1Glyph78.5
2Ovis2.5-9B66.4#1235
3Qwen 3 VL 8B Instruct24.3#1043
4GLM-4.1V-9B13.1#1134
5HiPPO12.1
6MiniCPM-V-2.610.3#3311
7Phi-3.5-Vision Instruct6.5#6279
8InternVL3-8B5.6#1795
9DeepSeek-VL25.6#3056
10LLaVA-OneVision-7B3.7#3002

No result here: #2 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=tablevision-compositional-arithmetic · How It Works · Data refreshed daily, snapshot 2026-11-22.