TableVista - Missing: leaderboard

Accuracy under the Missing rendering, which masks cells while the answer stays derivable, of all questions in TableVista (2026), a multimodal table reasoning benchmark in which models answer questions about rendered table images directly with thinking disabled, scored by exact match backed by a semantic-equivalence check. Percent; higher is better.

Metric: Accuracy (%) on the 3,000 TableVista questions under the Missing rendering (arbitrary cells masked while the answer stays derivable), direct-output prompt with thinking disabled (reasoning effort none for the GPT models), exact match with a GPT-5-mini semantic-equivalence check on exact-match failures; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 29 models tracked. Updated 7 Oct 2026.

Top models

#ModelSettingScoreOverall rank
1GPT-5.4Non-reasoning84.8#135
2Gemma 4 31B (IT)75.7#350
3Qwen 2.5 VL 72B Instruct71.3#841
4Qwen 2.5 VL 32B Instruct67.3#1062
5Llama 4 Maverick Instruct FP866.9#1023
6GPT-5.4 MiniNon-reasoning66.9#406
7Gemma 4 26B A4B (IT)65.6#556
8InternVL3.5-30B-A3B65.1#1267
9Llama 4 Scout Instruct64.9#1579
10Qwen 3.5 27BNon-reasoning63.8#375
11Qwen 3.5 122B A10BNon-reasoning63.3#346
12InternVL3.5-14B63#1504
13Qwen 2.5 VL 7B Instruct61.8#1902
14Gemma 3 27B (IT)61.1#1324
15Qwen 3 VL 30B A3B Instruct60.2#895

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=tablevista-missing · How It Works · Data refreshed daily, snapshot 2026-11-21.