mlsanityData to 14 Oct 2026

LongHarness: Advancing Agentic Reasoning over 100M-Token Corpora via Stateful Evidence Management

arXiv 2610.13075 ↗First posted 9 Oct 2026cs.AIcited by 0

On the map: LLM agents and tool use → Web, search, and research agents also fits: Agent memory systems

Abstract (arXiv)

Corpus-level long-context reasoning requires agents to acquire and integrate evidence across document collections that exceed their context windows. Existing agent harnesses enable incremental access to these corpora, but acquired evidence and evolving conclusions often remain entangled in a growing interaction history. This implicit state management can lead to two coupled failure modes: unbounded exploration, where retrieval expands the interaction history without resolving decision-relevant information gaps, and evidence-state drift, where previously acquired evidence or subsequent corrections fail to inform the current conclusion. We present LongHarness, a model-agnostic harness that makes the interaction between acquired evidence and evolving conclusions explicit. Its core mechanism, Evidence-Conclusion Coevolution, uses unresolved conclusions to guide exploration toward missing information, while newly acquired evidence strengthens or revises the conclusions that directed the search. LongHarness couples this process with Evidence Lifecycle Management, which keeps accumulated evidence current and traceable throughout the interaction. Together, these mechanisms focus retrieval on remaining information needs and keep conclusions aligned with current evidence. We evaluate LongHarness on OfficeQA Pro V1 and V2, AA-LCR, and LongBench v2 using GLM-5.3-Flash and Qwen3.8-Flash. Compared with direct reasoning using the same backbone, LongHarness improves GLM-5.3-Flash exact match on OfficeQA Pro from 15.8% to 60.2% on V1 and from 4.4% to 61.1% on V2. Against state-of-the-art agent harnesses, it surpasses Claude Code by 12.0 percentage points on V1 and Codex by 18.9 points on V2. With Qwen3.8-Flash, LongHarness reaches 61.7% on V1, outperforming Claude Code by 8.3 points and Codex by 12.0 points.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Web, search, and research agents, one of 16 directions in LLM agents and tool use. Builds on 38 of its 51 references that are in the map. First posted on arXiv 9 Oct 2026.

Builds on

38 of 51 references.

Closest ancestors · 2

Same direction, newest first

  1. 2025Self-Taught Agentic Long Context Understanding21 Feb
  2. 2024GraphReader: Building Graph-based Agent to Enhance Long-Context Abilities of Large Language Models20 Jun

Other references · 36

  1. 2022ReAct: Synergizing Reasoning and Acting in Language Modelscited by 8,695
  2. 2023Lost in the Middle: How Language Models Use Long Contextscited by 3,455
  3. 2024Gemini 1.5: Unlocking multimodal understanding across millions of tokens of contextcited by 3,224
  4. 2024RULER: What's the Real Context Size of Your Long-Context Language Models?cited by 1,346
  5. 2026DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligencecited by 1,040
  6. 2024LongBench v2: Towards Deeper Understanding and Reasoning on Realistic Long-context Multitaskscited by 379
  7. 2024Chain of Agents: Large Language Models Collaborating on Long-Context Taskscited by 206
  8. 2025Qwen2.5-1M Technical Reportcited by 164
  9. 2024A Human-Inspired Reading Agent with Gist Memory of Very Long Contextscited by 146
  10. 2026Agentic Harness Engineering: Observability-Driven Automatic Evolution of Coding-Agent Harnessescited by 142
  11. 2023Walking Down the Memory Maze: Beyond Context Limit through Interactive Readingcited by 133
  12. 2025Recursive Language Modelscited by 104
24 more
  1. 2025LongReason: A Synthetic Long-Context Reasoning Benchmark via Context Expansioncited by 50
  2. 2025QwenLong-L1: Towards Long-Context Large Reasoning Models with Reinforcement Learningcited by 50
  3. 2025LoongRL: Reinforcement Learning for Advanced Reasoning over Long Contextscited by 29
  4. 2025QwenLong-L1.5: Post-Training Recipe for Long-Context Reasoning and Memory Managementcited by 27
  5. 2025LongRoPE2: Near-Lossless LLM Context Window Scalingcited by 24
  6. 2025Beyond the Final Answer: Evaluating the Reasoning Trajectories of Tool-Augmented Agentscited by 24
  7. 2026LongRLVR: Long-Context Reinforcement Learning Requires Verifiable Context Rewardscited by 15
  8. 2026Prime Agent: A Self-Improving RLM Harnesscited by 13
  9. 2026Coding Agents are Effective Long-Context Processorscited by 12
  10. 2026LongR: Unleashing Long-Context Reasoning via Reinforcement Learning with Dense Utility Rewardscited by 10
  11. 2026Memory-R2: Fair Credit Assignment for Long-Horizon Memory-Augmented LLM Agentscited by 10
  12. 2026Recursive Experiential-Working Memory Evolution for Long-Horizon Agent Harnessescited by 9
  13. 2026Learning to Remember: End-to-End Training of Memory Agents for Long-Context Reasoningcited by 8
  14. 2026GoLongRL: Capability-Oriented Long Context Reinforcement Learning with Multitask Alignmentcited by 6
  15. 2026LEDGERMIND: Provenance-Constrained Multimodal Agentic Reasoning with a Structured Evidence Ledgercited by 5
  16. 2025DocPuzzle: A Process-Aware Benchmark for Evaluating Realistic Long-Context Reasoning Capabilitiescited by 4
  17. 2026ReContext: Recursive Evidence Replay as LLM Harness for Long-Context Reasoningcited by 4
  18. 2026OneDayAgent: Towards a Long-Horizon Harness for Autonomous Agentscited by 4
  19. 2026Incentivizing In-depth Reasoning over Long Contexts with Process Advantage Shapingcited by 3
  20. 2026SLIDERS: Systematic Reviews via Automated Evidence Synthesis and Reconciliationcited by 2
  21. 2026ACC: Compiling Agent Trajectories for Long-Context Trainingcited by 2
  22. 2026LongTraceRL: Learning Long-Context Reasoning from Search Agent Trajectories with Rubric Rewardscited by 2
  23. 2026ContextPilot: Teaching Agents for Proactive Context Management via Fine-grained RLcited by 2
  24. 2026Evidence-State Rewards for Long-Context Reasoningcited by 1

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026HMARS: A Hierarchical Multi-Agent Memory System for Long-Context Reasoning
  2. 2026Evidence Blindness in Direct Corpus Interaction: Persistent Navigation with AtlasNav
  3. 2026Rethinking Reasoning-Intensive Retrieval: Evaluating and Advancing Retrievers in Agentic Search Systems
  4. 2026Before Reasoning Can Fail: Pre-Evidence Procedural Failures in Agentic RAG
  5. 2026DocAtlas: Long-Document Understanding as Mutable-State Interaction
  6. 2026Retrieval-Infused Reasoning Sandbox: A Benchmark for Decoupling Retrieval and Reasoning Capabilities
  7. 2026Track, Rank, Crack: Epistemic Working Memory Scales Multi-Hop Reasoning in Language Agents
  8. 2026From Delivery to Stateful Exploration: Rethinking the Index for Agentic Search