Perplexity releases pplx-embed-v2-late multimodal embedding models

Perplexity has released pplx-embed-v2-late, two late-interaction embedding models designed to retrieve text, images, and pages through a shared embedding space for cross-model querying. Perplexity stated that both models achieve frontier performance.

Both models are publicly available on Hugging Face, with additional details shared in a [blog post](https://www.perplexity.ai/hub/blog/multimodal-embeddings-beyond-a-single-vector).