multimodal

Introducing the Multimodal Knowledge Agent benchmark

We’re launching Multimodal Knowledge Agent, a new benchmark to evaluate how agents reason across visually-complex multimodal tasks.