VisReason (Vision-Centric) - Board Reasoning: leaderboard

VisReason Board Reasoning is a structural-reasoning category of VisReason (2026): the model must infer the correct outcome of a board-game position shown as an image. The score is accuracy in percent, higher is better.

Metric: Accuracy (%) on the Board Reasoning category (structural reasoning: infer the correct outcome of a board-game position shown as an image) of VisReason, answers scored by format (regular-expression match, GPT-5-mini judge for open-ended answers, IoU above 0.5 for boxes), zero-shot with chain-of-thought prompt templates, one per answer format; rows not shaded gray in the paper use explicit reasoning (GPT-5 and Gemini 3 Pro at low reasoning effort); higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 20 models tracked. Updated 7 Oct 2026.

Top models

#ModelSettingScoreOverall rank
1Gemini 3 Pro (Preview)Low51.1#123
2GPT-5Low44.4#185
3GPT-5.2Thinking43.7#190
4GPT-5 MiniHigh42.2#373
5Qwen 3 VL 235B A22B (Thinking)30.4#462
6Qwen 3 VL 32B (Thinking)27.4#577
7O4 Mini26.7#372
8GPT-5 Nano22.2#912
9GPT-4o18.5#754
10Qwen 3 VL 30B A3B (Thinking)17#727
11MiMo-VL-7B-RL17#1160
12Gemini 2.5 Flash15.6#466
13Qwen 3 VL 8B (Thinking)14.1#924
14Qwen 3 VL 8B Instruct13.1#1043
15Qwen 2.5 VL 32B Instruct8.2#1065

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

Interactive version: theaggregate.ai/benchmark?slug=visreason-vision-centric-board-reasoning · How It Works · Data refreshed daily, snapshot 2026-11-22.