TableVista - Multi-Table: leaderboard

Accuracy on the questions spanning several tables (Web-style rendering) of 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 700 multi-table questions of the 3,000 TableVista questions, Web rendering, 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-reasoning61.3#135
2Qwen 2.5 VL 72B Instruct53.1#841
3Llama 4 Maverick Instruct FP852.4#1023
4Gemma 4 31B (IT)52.3#350
5Qwen 2.5 VL 32B Instruct49.1#1062
6Qwen 3.5 27BNon-reasoning48.6#375
7Llama 4 Scout Instruct48.1#1579
8Qwen 3.5 122B A10BNon-reasoning46.3#346
9InternVL3.5-30B-A3B43.3#1267
10Qwen 2.5 VL 7B Instruct42.4#1902
11Gemma 4 26B A4B (IT)42#556
12InternVL3.5-14B41.7#1504
13Qwen 3 VL 30B A3B Instruct41#895
14Qwen 3.6 35B A3BNon-reasoning41#481
15GPT-5.4 MiniNon-reasoning40#406

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

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