Building AI search in-house goes far beyond the prototype stage. Discover the hidden costs of building it yourself and how to accelerate the path to production without rebuilding the entire search stack.

Two 307M-parameter multilingual retrieval models that achieve state-of-the-art results on English general-domain retrieval (BEIR), long-document retrieval (MLDR), multilingual retrieval (MIRACL) and code retrieval (MTEB Code).

One 2-billion-parameter model reads text passages and scanned pages on the same relevance scale, from a single adapter and a single deployment.