Perplexity releases open-weight Embed v2 Context for text, image and document retrieval
Perplexity has released open-weight Embed v2 Context vision-language embedding models in 9B and 600M sizes for retrieval over text, images and PDFs.
Perplexity has released Embed v2 Context, a family of open-weight vision-language embedding models in two sizes, 9 billion parameters and 600 million parameters. They are built for retrieval over text, images and visual documents such as PDFs. The company describes the models as state-of-the-art for that work and has published a paper on them, according to Matthew Berman of Forward Future.
The models are meant for finding things. According to Berman, they are very good at locating text, images and documents. He says they are small enough to run locally on your own computer.
What changed
Perplexity released open-weight Embed v2 Context vision-language embedding models in 9B and 600M sizes.
Type: New capability · Applies to: Embed v2 Context
Why it matters
Open embedding models that handle visual documents could simplify building search over scanned or image-heavy company files.
What you can do
Teams building document search could test the 600M version against their current embedding model on visual documents.
Original source
We have no link to the original source for this story yet.
Where we saw it
Forward Future (Matthew Berman)
AI wrote this summary from the sources listed, and nobody reviewed it individually before publication. Check the original source before you rely on it. How it works