Benchmark

VQAv2

vision multimodal reasoning multimodal

VQAv2 is a balanced Visual Question Answering dataset that addresses language bias by providing complementary images for each question, forcing models to rely on visual understanding rather than language priors. It contains approximately twice the number of image-question pairs compared to the original VQA dataset.

语言EN
满分1
参评模型3

模型排名

名次 模型 机构 分数 来源
1 Pixtral Large Mistral AI 80.9 来源 ↗
2 Pixtral-12B Mistral AI 78.6 来源 ↗
3 Llama 3.2 90B Instruct Meta 78.1 来源 ↗