Search-o1: Agentic Search-Enhanced Large Reasoning Models

Large reasoning models (LRMs) like OpenAI-o1 have demonstrated impressivelong stepwise reasoning capabilities through large-scale reinforcementlearning. However, their extended reasoning processes often suffer fromknowledge insufficiency, leading to frequent uncertainties and potentialerrors. To address this limitation, we introduce Search-o1, aframework that enhances LRMs with an agentic retrieval-augmented generation(RAG) mechanism and a Reason-in-Documents module for refining retrieveddocuments. Search-o1 integrates an agentic search workflow into the reasoningprocess, enabling dynamic retrieval of external knowledge when LRMs encounteruncertain knowledge points. Additionally, due to the verbose nature ofretrieved documents, we design a separate Reason-in-Documents module to deeplyanalyze the retrieved information before injecting it into the reasoning chain,minimizing noise and preserving coherent reasoning flow. Extensive experimentson complex reasoning tasks in science, mathematics, and coding, as well as sixopen-domain QA benchmarks, demonstrate the strong performance of Search-o1.This approach enhances the trustworthiness and applicability of LRMs in complexreasoning tasks, paving the way for more reliable and versatile intelligentsystems. The code is available athttps://github.com/sunnynexus/Search-o1.