AutoResearchBench is a new benchmark showing top AI agents achieve under 10% success on complex scientific literature discovery tasks that demand deep comprehension and open-ended search.
super hub Canonical reference
Search-R1: Training LLMs to Reason and Leverage Search Engines with Reinforcement Learning
Canonical reference. 78% of citing Pith papers cite this work as background.
125
Pith papers citing it
Background
78% of classified citations
abstract
Efficiently acquiring external knowledge and up-to-date information is essential for effective reasoning and text generation in large language models (LLMs). Prompting advanced LLMs with reasoning capabilities to use search engines during inference is often suboptimal, as the LLM might not fully possess the capability on how to interact optimally with the search engine. This paper introduces Search-R1, an extension of reinforcement learning (RL) for reasoning frameworks where the LLM learns to autonomously generate (multiple) search queries during step-by-step reasoning with real-time retrieval. Search-R1 optimizes LLM reasoning trajectories with multi-turn search interactions, leveraging retrieved token masking for stable RL training and a simple outcome-based reward function. Experiments on seven question-answering datasets show that Search-R1 improves performance by 41% (Qwen2.5-7B) and 20% (Qwen2.5-3B) over various RAG baselines under the same setting. This paper further provides empirical insights into RL optimization methods, LLM choices, and response length dynamics in retrieval-augmented reasoning. The code and model checkpoints are available at https://github.com/PeterGriffinJin/Search-R1.
hub tools
citation-role summary
background 42
method 3
baseline 2
dataset 1
extension 1
other 1
citation-polarity summary
claims ledger
- abstract Efficiently acquiring external knowledge and up-to-date information is essential for effective reasoning and text generation in large language models (LLMs). Prompting advanced LLMs with reasoning capabilities to use search engines during inference is often suboptimal, as the LLM might not fully possess the capability on how to interact optimally with the search engine. This paper introduces Search-R1, an extension of reinforcement learning (RL) for reasoning frameworks where the LLM learns to autonomously generate (multiple) search queries during step-by-step reasoning with real-time retrieva
authors
co-cited works
representative citing papers
Co-ReAct adds step-level rubric guidance to ReAct agents via a GRPO-trained generator using list-wise ranking rewards, yielding consistent gains on DeepResearchBench and SQA-CS-V2.
SVFSearch is the first open benchmark for short-video frame search in the Chinese gaming domain, providing a frozen retrieval environment and showing performance gaps of 13-29 points between direct QA models, practical agents, and oracle knowledge.
AstraFlow decouples RL components into autonomous dataflow services to natively support multi-policy agentic LLM training, elastic scaling, and cross-region execution with 2.7x speedup on math, code, search, and AgentBench workloads.
PyRAG turns multi-hop reasoning into executable Python code over retrieval tools for explicit, verifiable step-by-step RAG.
ActGuide-RL uses human action data as plan-style guidance in mixed-policy RL to overcome exploration barriers in LLM agents, matching SFT+RL performance on search benchmarks without cold-start training.
CuSearch reallocates rollout budget in RLVR toward deeper-search trajectories as a proxy for retrieval supervision density, yielding up to 11.8 exact-match gains over uniform GRPO sampling on ZeroSearch.
Evolving-RL jointly optimizes experience extraction and utilization in LLM agents via RL with separate evaluation signals, delivering up to 98.7% relative gains on out-of-distribution tasks in ALFWorld and Mind2Web.
DeepRefine refines agent-compiled knowledge bases via multi-turn abductive diagnosis and RL training with a GBD reward, yielding consistent downstream task gains.
LLM agents encode tool necessity in pre-generation hidden states with high linear decodability (AUROC 0.89-0.96); Probe&Prefill uses this to reduce tool calls 48% with 1.7% accuracy loss.
SearchSkill improves exact match scores and retrieval efficiency on open-domain QA by conditioning LLM actions on skills from an evolving SkillBank updated from failure patterns via two-stage SFT.
AHD Agent trains a 4B-parameter LLM via agentic RL to actively use tools for automatic heuristic design, matching or exceeding larger baselines across eight domains with fewer evaluations.
AgentForesight introduces an online auditor model that predicts decisive errors in multi-agent trajectories at the earliest step using a coarse-to-fine reinforcement learning recipe on a new curated dataset AFTraj-2K.
The cancellation hypothesis shows how rollout-level rewards produce token-level credit assignment in critic-free RL through cancellation of opposing signals on shared tokens, with empirical support and batching interventions that enhance performance.
AIDA is the first end-to-end autonomous agent that combines a domain-specific language with Pareto-guided reinforcement learning to discover insights from complex business data.
A VOI-based controller for dual inference budgets improves multi-hop QA performance by prioritizing search actions and selectively finalizing answers.
Introduces the SRA paradigm and SRA-Bench benchmark showing retrieval-based skill augmentation improves agent performance but skill incorporation remains a bottleneck regardless of retrieval quality.
MAGEO is a multi-agent system that distills validated editing patterns into reusable optimization skills for generative engines, outperforming heuristic baselines on visibility and fidelity via a new benchmark and evaluation protocol.
LAnR unifies retrieval-augmented generation inside a single LLM by deriving dense retrieval vectors from a [PRED] token's hidden states and using entropy to adaptively stop retrieval, outperforming prior RAG on six QA benchmarks with better efficiency.
Deep-Reporter introduces a unified agentic framework for grounded multimodal long-form generation via multimodal search, checklist-guided synthesis, and recurrent context management, plus the M2LongBench benchmark.
Agentic search narrows the gap between dense RAG and GraphRAG but does not remove GraphRAG's advantage on complex multi-hop reasoning.
SDRL trains LLMs via self-generated multi-path debates and joint optimization of standalone plus debate-conditioned responses to boost both single-model reasoning and multi-agent debate performance.
Temp-R1 uses reverse curriculum reinforcement learning to train an autonomous agent that achieves state-of-the-art results on temporal KGQA benchmarks by developing sophisticated reasoning on hard questions first.
VideoDR is a new benchmark for open-web video deep research that tests multimodal models on cross-frame visual anchor extraction, interactive retrieval, and multi-hop reasoning over joint video-web evidence.
citing papers explorer
-
AutoResearchBench: Benchmarking AI Agents on Complex Scientific Literature Discovery
AutoResearchBench is a new benchmark showing top AI agents achieve under 10% success on complex scientific literature discovery tasks that demand deep comprehension and open-ended search.
-
Co-ReAct: Rubrics as Step-Level Collaborators for ReAct Agents
Co-ReAct adds step-level rubric guidance to ReAct agents via a GRPO-trained generator using list-wise ranking rewards, yielding consistent gains on DeepResearchBench and SQA-CS-V2.
-
SVFSearch: A Multimodal Knowledge-Intensive Benchmark for Short-Video Frame Search in the Gaming Vertical Domain
SVFSearch is the first open benchmark for short-video frame search in the Chinese gaming domain, providing a frozen retrieval environment and showing performance gaps of 13-29 points between direct QA models, practical agents, and oracle knowledge.
-
AstraFlow: Dataflow-Oriented Reinforcement Learning for Agentic LLMs
AstraFlow decouples RL components into autonomous dataflow services to natively support multi-policy agentic LLM training, elastic scaling, and cross-region execution with 2.7x speedup on math, code, search, and AgentBench workloads.
-
Retrieval is Cheap, Show Me the Code: Executable Multi-Hop Reasoning for Retrieval-Augmented Generation
PyRAG turns multi-hop reasoning into executable Python code over retrieval tools for explicit, verifiable step-by-step RAG.
-
Learning Agentic Policy from Action Guidance
ActGuide-RL uses human action data as plan-style guidance in mixed-policy RL to overcome exploration barriers in LLM agents, matching SFT+RL performance on search benchmarks without cold-start training.
-
CuSearch: Curriculum Rollout Sampling via Search Depth for Agentic RAG
CuSearch reallocates rollout budget in RLVR toward deeper-search trajectories as a proxy for retrieval supervision density, yielding up to 11.8 exact-match gains over uniform GRPO sampling on ZeroSearch.
-
Evolving-RL: End-to-End Optimization of Experience-Driven Self-Evolving Capability within Agents
Evolving-RL jointly optimizes experience extraction and utilization in LLM agents via RL with separate evaluation signals, delivering up to 98.7% relative gains on out-of-distribution tasks in ALFWorld and Mind2Web.
-
DeepRefine: Agent-Compiled Knowledge Refinement via Reinforcement Learning
DeepRefine refines agent-compiled knowledge bases via multi-turn abductive diagnosis and RL training with a GBD reward, yielding consistent downstream task gains.
-
LLM Agents Already Know When to Call Tools -- Even Without Reasoning
LLM agents encode tool necessity in pre-generation hidden states with high linear decodability (AUROC 0.89-0.96); Probe&Prefill uses this to reduce tool calls 48% with 1.7% accuracy loss.
-
SearchSkill: Teaching LLMs to Use Search Tools with Evolving Skill Banks
SearchSkill improves exact match scores and retrieval efficiency on open-domain QA by conditioning LLM actions on skills from an evolving SkillBank updated from failure patterns via two-stage SFT.
-
AHD Agent: Agentic Reinforcement Learning for Automatic Heuristic Design
AHD Agent trains a 4B-parameter LLM via agentic RL to actively use tools for automatic heuristic design, matching or exceeding larger baselines across eight domains with fewer evaluations.
-
AgentForesight: Online Auditing for Early Failure Prediction in Multi-Agent Systems
AgentForesight introduces an online auditor model that predicts decisive errors in multi-agent trajectories at the earliest step using a coarse-to-fine reinforcement learning recipe on a new curated dataset AFTraj-2K.
-
The Cancellation Hypothesis in Critic-Free RL: From Outcome Rewards to Token Credits
The cancellation hypothesis shows how rollout-level rewards produce token-level credit assignment in critic-free RL through cancellation of opposing signals on shared tokens, with empirical support and batching interventions that enhance performance.
-
Towards Autonomous Business Intelligence via Data-to-Insight Discovery Agent
AIDA is the first end-to-end autonomous agent that combines a domain-specific language with Pareto-guided reinforcement learning to discover insights from complex business data.
-
Inference-Time Budget Control for LLM Search Agents
A VOI-based controller for dual inference budgets improves multi-hop QA performance by prioritizing search actions and selectively finalizing answers.
-
Skill Retrieval Augmentation for Agentic AI
Introduces the SRA paradigm and SRA-Bench benchmark showing retrieval-based skill augmentation improves agent performance but skill incorporation remains a bottleneck regardless of retrieval quality.
-
From Experience to Skill: Multi-Agent Generative Engine Optimization via Reusable Strategy Learning
MAGEO is a multi-agent system that distills validated editing patterns into reusable optimization skills for generative engines, outperforming heuristic baselines on visibility and fidelity via a new benchmark and evaluation protocol.
-
Latent Abstraction for Retrieval-Augmented Generation
LAnR unifies retrieval-augmented generation inside a single LLM by deriving dense retrieval vectors from a [PRED] token's hidden states and using entropy to adaptively stop retrieval, outperforming prior RAG on six QA benchmarks with better efficiency.
-
Deep-Reporter: Deep Research for Grounded Multimodal Long-Form Generation
Deep-Reporter introduces a unified agentic framework for grounded multimodal long-form generation via multimodal search, checklist-guided synthesis, and recurrent context management, plus the M2LongBench benchmark.
-
Do We Still Need GraphRAG? Benchmarking RAG and GraphRAG for Agentic Search Systems
Agentic search narrows the gap between dense RAG and GraphRAG but does not remove GraphRAG's advantage on complex multi-hop reasoning.
-
Learning from Self-Debate: Preparing Reasoning Models for Multi-Agent Debate
SDRL trains LLMs via self-generated multi-path debates and joint optimization of standalone plus debate-conditioned responses to boost both single-model reasoning and multi-agent debate performance.
-
Temp-R1: A Unified Autonomous Agent for Complex Temporal KGQA via Reverse Curriculum Reinforcement Learning
Temp-R1 uses reverse curriculum reinforcement learning to train an autonomous agent that achieves state-of-the-art results on temporal KGQA benchmarks by developing sophisticated reasoning on hard questions first.
-
Watching, Reasoning, and Searching: A Video Deep Research Benchmark on Open Web for Agentic Video Reasoning
VideoDR is a new benchmark for open-web video deep research that tests multimodal models on cross-frame visual anchor extraction, interactive retrieval, and multi-hop reasoning over joint video-web evidence.
-
Training Multi-Image Vision Agents via End2End Reinforcement Learning
IMAgent trains a multi-image vision agent via pure end-to-end RL with visual reflection tools and a two-layer motion trajectory masking strategy, reaching SOTA on single- and multi-image benchmarks while revealing tool-use effects on attention.
-
Q-RAG: Long Context Multi-step Retrieval via Value-based Embedder Training
Q-RAG trains embedders via RL for multi-step retrieval and reports state-of-the-art results on BabiLong and RULER benchmarks for contexts up to 10M tokens.
-
MMSearch-R1: Incentivizing LMMs to Search
MMSearch-R1 uses reinforcement learning to train multimodal models for on-demand multi-turn internet search with image and text tools, outperforming same-size RAG baselines and matching larger ones while cutting search calls by over 30%.
-
Group-in-Group Policy Optimization for LLM Agent Training
GiGPO adds a hierarchical grouping mechanism to group-based RL so that LLM agents receive both global trajectory and local step-level credit signals, yielding >12% gains on ALFWorld and >9% on WebShop over GRPO while keeping the same rollout and memory footprint.
-
ReflectMT: Internalizing Reflection for Efficient and High-Quality Machine Translation
ReflectMT internalizes reflection via two-stage RL to enable direct high-quality machine translation that outperforms explicit reasoning models like DeepSeek-R1 on WMT24 while using 94% fewer tokens.
-
Freshness-Aware Prioritized Experience Replay for LLM/VLM Reinforcement Learning
Freshness-Aware PER augments prioritized experience replay with exponential age decay based on effective sample size to enable successful reuse of trajectories in LLM and VLM reinforcement learning, outperforming on-policy baselines on agentic tasks.
-
Does RL Expand the Capability Boundary of LLM Agents? A PASS@(k,T) Analysis
RL expands the capability boundary of LLM agents on compositional tool-use tasks, shown by non-converging pass curves at large k with increasing T, while SFT regresses it and the effect is absent on simpler tasks.
-
AdversarialCoT: Single-Document Retrieval Poisoning for LLM Reasoning
A single query-specific poisoned document, built by extracting and iteratively refining an adversarial chain-of-thought, can substantially degrade reasoning accuracy in retrieval-augmented LLM systems.
-
Controllable and Verifiable Tool-Use Data Synthesis for Agentic Reinforcement Learning
COVERT generates verifiable synthetic tool-use environments for RL by validated trajectory synthesis and oracle-preserving augmentations, improving tool-use accuracy on BFCL v3 and ACEBench while remaining complementary to SFT.
-
VISOR: Agentic Visual Retrieval-Augmented Generation via Iterative Search and Over-horizon Reasoning
VISOR is a unified agentic VRAG framework with Evidence Space structuring, visual action evaluation/correction, and dynamic sliding-window trajectories trained via GRPO-based RL that achieves SOTA performance on long-horizon visual reasoning benchmarks.
-
Retrieval Augmented Conversational Recommendation with Reinforcement Learning
RAR retrieves candidate items from a 300k-movie corpus then uses LLM generation with RL feedback to produce context-aware recommendations that outperform baselines on benchmarks.
-
DeferMem: Query-Time Evidence Distillation via Reinforcement Learning for Long-Term Memory QA
DeferMem decouples memory QA into high-recall retrieval and RL-based query-conditioned evidence distillation, outperforming baselines on LoCoMo and LongMemEval-S with highest accuracy, fastest runtime, and zero API token cost.
-
Maestro: Reinforcement Learning to Orchestrate Hierarchical Model-Skill Ensembles
Maestro uses outcome-based RL to train a lightweight policy that orchestrates ensembles of frozen expert models and skills, reporting 70.1% average accuracy across ten multimodal benchmarks and outperforming GPT-5 and Gemini-2.5-Pro while generalizing to unseen components.
-
Efficient Agentic Reasoning Through Self-Regulated Simulative Planning
SR²AM achieves competitive Pass@1 accuracy on diverse tasks with 25.8-95.3% fewer reasoning tokens than much larger models by using self-regulated simulative planning trained via supervised learning and RL.
-
What and When to Distill: Selective Hindsight Distillation for Multi-Turn Agents
SERL selectively reweights learning using task success and environment feedback to reach 90.0% success on ALFWorld and 80.1% on WebShop, outperforming RL and distillation baselines.
-
Harnessing LLM Agents with Skill Programs
HASP upgrades textual skills into executable Program Functions that intervene in LLM agent loops at inference, post-training, or self-evolution, delivering 25% gains over ReAct and 30.4% over Search-R1 on reasoning benchmarks.
-
RubricEM: Meta-RL with Rubric-guided Policy Decomposition beyond Verifiable Rewards
RubricEM uses rubric-guided stagewise policy decomposition and reflection-based meta-policy evolution to improve long-horizon research agents beyond verifiable rewards.
-
ComplexMCP: Evaluation of LLM Agents in Dynamic, Interdependent, and Large-Scale Tool Sandbox
ComplexMCP benchmark shows top LLM agents achieve under 60% success on dynamic interdependent tool tasks versus 90% for humans, due to tool retrieval saturation, over-confidence, and strategic defeatism.
-
PruneTIR: Inference-Time Tool Call Pruning for Effective yet Efficient Tool-Integrated Reasoning
PruneTIR prunes erroneous tool-call trajectories during LLM inference via three trigger-based components to raise Pass@1 accuracy and efficiency while shortening context.
-
TIDE-Bench: Task-Aware and Diagnostic Evaluation of Tool-Integrated Reasoning
TIDE-Bench is a new benchmark for tool-integrated reasoning that combines diverse tasks, multi-aspect metrics covering answer quality, process reliability, efficiency and cost, plus filtered challenging test sets.
-
PiCA: Pivot-Based Credit Assignment for Search Agentic Reinforcement Learning
PiCA uses pivot-based potential rewards derived from historical sub-queries to supply trajectory-aware step guidance in agentic RL, delivering 15% gains on QA benchmarks for 3B/7B models.
-
Beyond Thinking: Imagining in 360$^\circ$ for Humanoid Visual Search
Imagining in 360° decouples visual search into a single-step probabilistic semantic layout predictor and an actor, removing the need for multi-turn CoT reasoning and trajectory annotations while improving efficiency in 360° environments.
-
SkillMaster: Toward Autonomous Skill Mastery in LLM Agents
SkillMaster enables LLM agents to autonomously develop skills via trajectory review, counterfactual evaluation, and DualAdv-GRPO training, boosting success rates by 8.8% on ALFWorld and 9.3% on WebShop.
-
Iterative Critique-and-Routing Controller for Multi-Agent Systems with Heterogeneous LLMs
A critique-and-routing controller cast as a finite-horizon MDP with policy-gradient optimization outperforms one-shot routing baselines on reasoning benchmarks while using the strongest agent for under 25% of calls.
-
AIPO: Learning to Reason from Active Interaction
AIPO adds active multi-agent consultation (Verify, Knowledge, Reasoning agents) plus custom importance sampling to RLVR training so LLMs expand their reasoning boundary and then operate without the agents.
-
A$^2$TGPO: Agentic Turn-Group Policy Optimization with Adaptive Turn-level Clipping
A²TGPO improves RL policy optimization for multi-turn agentic LLMs by normalizing information gain within same-depth turn groups, rescaling cumulative advantages by sqrt of term count, and modulating clipping ranges per turn's normalized IG.