TRUSTED BY FRONTIER AI LABS
We shape how models learn. Data is not the bottleneck. How experience reaches the model is.
Vetto designs learning experiences, feedback systems, and evaluations that directly improve real-world model capability.
Plot Twist Bench
Do frontier models actually understand video? Plot Twist Bench is a video perception benchmark of 48 curated multiple-choice questions, made by Vetto.
Everyone accepts that a language model autocompletes text. Fewer people have noticed that a vision-language model autocompletes video: asked what happened in a clip, it answers with what usually happens. The only footage that separates watching from recognising is footage where the usual thing does not happen, and that is not on the internet, so we filmed it.
