MIRACL
MIRACL measures how well a system retrieves relevant Wikipedia passages when the query and the documents are in the same language, across 18 languages. Human annotators judged which passages were relevant to each query, giving a high-quality basis for comparing multilingual retrieval models.
It has become a standard reference for anyone building retrieval for non-English audiences, where results from English-only benchmarks often fail to carry over. The benchmark is released under Apache 2.0, so it is straightforward to use in commercial work.
Related sources
BEIR
A collection of 18 information retrieval datasets spanning ad-hoc web search, question answering, fact verification, and duplicate question retrieval. Built by aggregating existing datasets, some originally created for other tasks and converted to retrieval format. Tests how well retrieval models generalise to unseen domains without fine-tuning, which matters for real-world RAG.
BRIGHT
A reasoning-intensive retrieval benchmark. It requires genuine reasoning to connect queries with relevant documents rather than keyword or semantic overlap, revealing model weaknesses that BEIR misses.
CoIR (Code Retrieval Benchmark)
A code retrieval benchmark spanning diverse programming tasks and languages. Built for evaluating how well models retrieve relevant source code, which matters if you are building RAG over codebases.
LongEmbed
A retrieval benchmark focused on very long documents, with an average length above 5,500 words. It addresses a gap in standard benchmarks, which mostly use short documents of at most 512 tokens (the chunks of text a model reads at once).