Perplexity says its agents search real sandbox checkouts with grep and standard file tools instead of broad semantic searches. On an internal benchmark of 150 financial-document questions, this approach reached 0.627 answer accuracy versus about 0.37 for semantic search.
Part of a thread · earlier post
Because each sandbox holds a real checkout, agents search it with grep and standard file tools instead of relying on broad semantic searches.
On an internal benchmark of 150 financial-document questions, this reached 0.627 answer accuracy versus about 0.37 for semantic search.
Source: Perplexity · x.comPublished · added here
