Mistralのエージェント型検索の仕組みと、正答率が3倍に上がる理由

"Agentic Search builds on this index by giving the model five tools that resemble familiar file-system operations:"／"The index identifies likely sources; Agentic Search determines what to inspect within and across them."／"These tools do not require fine-tuning or model-specific training. As models get better at reasoning and tool use, retrievals get better without infrastructure changes. This is a key property: retrieval quality scales with model capability instead of being capped by your chunking strategy."／"We observe the same performance patterns across first- and third-party models, which indicates that Agentic Search is model-agnostic, and that search quality should improve with new models."（How Agentic Search works／More relevant results, faster 節）— Mistral AI 公式ブログより

"Agentic Search introduces a multi-step retrieval loop for finding, inspecting, and verifying information across data sources, wherever it is stored."／"without crossing your isolation boundaries in the cloud or on-premises."／"Mistral Search Toolkit provides open modules for ingesting, embedding, and indexing critical and complex data in the cloud or on-premises."／"It accumulates in financial filings, legal contracts, internal resources, and government records–long, dense documents that traditional search methods can't navigate effectively."／"It creates a local index for your own corpus using a default configuration, so you can try Agentic Search without needing to be a search expert."（冒頭／How Agentic Search works／Data creates competitive advantage／Getting started 節）— Mistral AI 公式ブログより

"Traditional, one-shot RAG retrieves a fixed set of text chunks and asks a model to answer in a single pass. This works when the answer appears in one of the top results, but falters when the model must navigate a long report, follow references, compare multiple documents, or verify the underlying evidence."／"The model must answer from the chunks selected during the initial retrieval, even when they are incomplete or not relevant."／"It cannot decide that it needs a different document, another section, or more context before responding"／"an index may find the correct document but cannot open it, navigate to the table, read the surrounding context, or verify the answer."（Traditional RAG falls short 節）— Mistral AI 公式ブログより

"Many questions need more than one retrieval pass to get the correct answer. The model may need to refine its search, inspect a promising document, follow a reference, compare multiple sources, keep track of what it has seen, and try a new route when the first results are insufficient. One-shot RAG provides no way to take these next steps."（Traditional RAG falls short 節）— Mistral AI 公式ブログより

"We benchmarked Agentic Search on two industry-standard evaluations, using the out-of-the-box Mistral Search Toolkit stack: default chunking, default ranking, no tuning. These results are floors, not ceilings"／"we tested two models using the Mistral Search Toolkit: Mistral Medium 3.5 (MM 3.5) and Z.ai GLM-5.2 (GLM-5.2), showcasing performance of a smaller model (MM 3.5) and a larger model (GLM-5.2)."／"FinanceBench (Islam et al., 2023) tests financial question-answering over 368 SEC filings (10-K / 10-Q / 8-K), averaging ~147 pages each, ~53,900 pages total: long, table-heavy financial documents. Answers scored by an LLM judge calibrated against human labels."／"Agentic Search delivers to 3x correctness on financial filings, from 26.7% to 86%, based on FinanceBench."／"The search-only Agentic loop is the biggest quality lever."／"Because models can search iteratively, they can recover from weak first results, refine queries, and use the index as an active tool."（冒頭／More relevant results, faster／FinanceBench 節）— Mistral AI 公式ブログより

"This means a targeted drill-in search beats repeated broad search in complex documents."／"The retrieval tools are not additional overhead"／"they replace wasted search retries with precise navigation."／"Across FinanceBench, adding navigation retrieval tools improves latency: p90 drops"／"In general, we see the search-only loops conduct repeated broad searches, while navigation helps the model identify evidence more quickly."（FinanceBench 節）— Mistral AI 公式ブログより

"On table-heavy, multi-doc questions of the OfficeQA Pro benchmark, we measure a +45.6 point gain (6.3% to 51.9%)."／"OfficeQA Pro is a verifiable numeric benchmark over historical U.S. Treasury Bulletins: scanned, table-heavy government-finance PDFs across a 696-document, ~89,000-page corpus."／"We report the first pass for the 133-question"／"OfficeQA Pro has numeric answers, scanned PDFs, and deep table lookups."／"The harder the benchmark, the more important the retrieval loop becomes."／"OfficeQA Pro is built around numeric answers in scanned, table-heavy documents. One-shot RAG barely gets started, while the agentic loop allows the model to search iteratively, inspect evidence, and deliver substantial accuracy improvements."（OfficeQA Pro 節）— Mistral AI 公式ブログより

"The tooling stack drives substantial impact on document intelligence and search performance."／"GLM-5.2 scores 41.4% on OfficeQA Pro with the Claude Code harness, compared with 51.9% on the Mistral harness"（OfficeQA Pro 節）— Mistral AI 公式ブログより

"Filings, contracts, manuals, technical specifications, and reports where the answer may appear on a particular page or in a specific table, clause, figure, or footnote."／"Research that requires the model to find, compare, or reconcile evidence from several documents before reaching an answer."／"Financial figures, legal clauses, regulatory references, and operational data, where the response can be referenced in a stable and specific document location."／"Financial statements, government records, and scanned PDFs where meaning depends on rows, columns, page position, or surrounding context"（Use Agentic Search for 節）— Mistral AI 公式ブログより

"Short, clean documents where the answer is likely to appear in one of the first retrieved chunks."／"Keyword or semantic lookups that need to return relevant passages without reasoning over or navigating through them."／"Use cases where the likely source and location of the answer are known in advance and additional retrieval steps are unlikely to improve the result."／"One-shot RAG is often sufficient for these searches. Add Agentic Search when questions require the model to move beyond the initial results and investigate the source material. A well-configured index remains the right foundation in both cases."（Indexed retrieval is the right starting point for 節）— Mistral AI 公式ブログより