Searchplex ID Retrieval terminology registry

Artifact Benchmark

BrowseComp-Plus

Canonical ID https://id.searchplex.net/browsecomp-plus/

A controlled benchmark for evaluating deep-research agents and retrievers, derived from BrowseComp and built around a fixed curated corpus with supporting documents and hard negatives.

Scope

BrowseComp-Plus is designed to make evaluation of deep-research agents and retrieval methods more fair and transparent by controlling the document corpus and evidence. It is related to BrowseComp but should not be treated as identical to live-web browsing evaluation.

Observed terms

  • BCP academic
  • BrowseComp Plus academic

Evaluates

Not equivalent to

  • BrowseComp-Plus is not BEIR. BrowseComp-Plus evaluates deep-research agents and retrieval methods over BrowseComp-derived tasks; BEIR is an information-retrieval benchmark suite.
  • BrowseComp-Plus is not BrowseComp. BrowseComp-Plus derives from BrowseComp but uses a fixed curated corpus and supporting evidence; BrowseComp evaluates browsing agents under live web browsing conditions.
  • BrowseComp-Plus is not MTEB. BrowseComp-Plus evaluates deep-research agents and retrievers; MTEB evaluates text embedding models across multiple tasks.

Learn more

ID