KnowHal - Positive Questions: leaderboard

Metric: Accuracy (%; mean over the entity, attribute, relation and knowledge dimensions on questions about valid visual evidence or facts; 1,800 entity-image samples in 10 domains with paired positive and false-premise negative questions; positive answers scored by standard VQA answer normalization and soft matching, negative answers by a GPT-4o-mini judge (option matching for relation questions); zero-shot). Source: arxiv.org. Saturation forecast: Around December 2026. 14 models tracked.

Top models

#ModelScore
1Gemini 3 Flash (Preview)86.6
2Seed 2.0 Mini85
3Kimi K2.579.2
4GLM-4.6V77.1
5Qwen 3 VL 32B Instruct76.9
6Qwen 2.5 VL 72B Instruct74.9
7GPT-5.6 Sol73.6
8GPT-5.267.9
9InternVL3.5-8B54.1

Interactive version: theaggregate.ai/benchmark?slug=knowhal-positive-questions · How It Works · Data refreshed daily, snapshot 2026-09-29.