Benchmark

BrowseComp Long Context 256k

BrowseComp is a benchmark for measuring the ability of agents to browse the web, comprising 1,266 questions that require persistently navigating the…

Modality
text
Categories
reasoning, search
Openness
unknown
Source
llm_stats
Reported scores
2

Reported scores

llm_stats

ModelOrganizationReported valueReportedEvidence
GPT-5OpenAI0.8882025-08-07evidence
GPT-5.2OpenAI0.8982025-12-11evidence

Scores are partitioned by the source that reported them and are never merged into a single cross-source ranking, because the sources measure different things and say so.