ARXIV:2603.09022 · AGENTS · SUBMITTED 02 APR · 02:30 UTC · FRESHNESS STALE

VerifiedSource: PDF linkedPartialPaperPack: 3 of 4 citation fields filledMissingMissing fields: authorsPartialProof: unverified proof status

MEMO: Memory-Augmented Model Context Optimization for Robust Multi-Turn Multi-Agent LLM Games

arXiv

MEMO optimizes context in multi-agent LLM games to improve performance and stability.

Blocked on Code›Score3.0Evidence unverified

Opportunity summary

Pain MEMO optimizes context in multi-agent LLM games to improve performance and stability.

Evidence 0 refs | 0 sources | 17% coverage

Blocker Evidence unverified

Open Build Read PDF Signal Canvas Track

PROBLEM

MEMO optimizes context in multi-agent LLM games to improve performance and stability. In long-horizon interactions, small early deviations compound across turns and are amplified by multi-agent coupling.

METHOD

Full abstract

Multi-turn, multi-agent LLM game evaluations often exhibit substantial run-to-run variance. In long-horizon interactions, small early deviations compound across turns and are amplified by multi-agent coupling. This biases win rate estimates and makes rankings unreliable across repeated tournaments. Prompt choice worsens this further by producing different effective policies. We address both instability and underperformance with MEMO (Memory-augmented MOdel context optimization), a self-play framework that optimizes inference-time context by coupling retention and exploration. Retention maintains a persistent memory bank that stores structured insights from self-play trajectories and injects them as priors during later play. Exploration runs tournament-style prompt evolution with uncertainty-aware selection via TrueSkill, and uses prioritized replay to revisit rare and decisive states. Across five text-based games, MEMO raises mean win rate from 25.1% to 49.5% for GPT-4o-mini and from 20.9% to 44.3% for Qwen-2.5-7B-Instruct, using $2,000$ self-play games per task. Run-to-run variance also drops, giving more stable rankings across prompt variations. These results suggest that multi-agent LLM game performance and robustness have substantial room for improvement through context optimization. MEMO achieves the largest gains in negotiation and imperfect-information games, while RL remains more effective in perfect-information settings.

RESULT

ScienceToStartup currently rates this 3.0/10 on the public viability pass. These results suggest that multi-agent LLM game performance and robustness have substantial room for improvement through context optimization.

WHY NOW

Agents moved forward this cycle; last verified April 2026. Public score 3.0/10.

Continue into Read for claims, analysis, references, and neighboring papers.

Opportunity summary

Score3.0

PainMEMO optimizes context in multi-agent LLM games to improve performance and stability.

Evidence0 refs | 0 sources | 17% coverage

Blockermissing authors

Analysis summary

MEMO optimizes context in multi-agent LLM games to improve performance and stability.

VerifiedSource: PDF linkedPartialPaperPack: 3 of 4 citation fields filledMissingMissing fields: authorsPartialProof: unverified proof status

{ "contract_version": "paper-r2", "paper_id": "f9e1afcb-ebd6-4458-8a0c-fea30eb7cc05", "arxiv_id": "2603.09022", "canonical_route": "/paper/memo-memory-augmented-model-context-optimization-for-robust-multi-turn-multi-agent-llm-games", "active_tab": "synced from current hash by the drawer client", "selected_artifact": "memo-memory-augmented-model-context-optimization-for-robust-multi-turn-multi-agent-llm-games", "endpoints": { "paper_pack": "/api/v1/paper/memo-memory-augmented-model-context-optimization-for-robust-multi-turn-multi-agent-llm-games/paper-pack", "build_passport": "/api/v1/paper/memo-memory-augmented-model-context-optimization-for-robust-multi-turn-multi-agent-llm-games/build-passport", "mcp_resource": "sciencetostartup://surfaces/paper-workspace" } }

{ "surface": "paper", "mode": "paper", "query": "MEMO: Memory-Augmented Model Context Optimization for Robust Multi-Turn Multi-Agent LLM Games", "normalized_query": "2603.09022", "route": "/paper/memo-memory-augmented-model-context-optimization-for-robust-multi-turn-multi-agent-llm-games", "paper_ref": "memo-memory-augmented-model-context-optimization-for-robust-multi-turn-multi-agent-llm-games", "topic_slug": null, "benchmark_ref": null, "dataset_ref": null }

{ "@context": "https://schema.org", "@graph": [ { "@type": "WebPage", "@id": "https://sciencetostartup.com/paper/memo-memory-augmented-model-context-optimization-for-robust-multi-turn-multi-agent-llm-games#webpage", "url": "https://sciencetostartup.com/paper/memo-memory-augmented-model-context-optimization-for-robust-multi-turn-multi-agent-llm-games", "name": "MEMO: Memory-Augmented Model Context Optimization for Robust Multi-Turn Multi-Agent LLM Games", "description": "MEMO optimizes context in multi-agent LLM games to improve performance and stability.", "isPartOf": { "@id": "https://sciencetostartup.com/#website" } }, { "@type": "ScholarlyArticle", "@id": "https://sciencetostartup.com/paper/memo-memory-augmented-model-context-optimization-for-robust-multi-turn-multi-agent-llm-games#scholarlyArticle", "headline": "MEMO: Memory-Augmented Model Context Optimization for Robust Multi-Turn Multi-Agent LLM Games", "description": "MEMO optimizes context in multi-agent LLM games to improve performance and stability.", "url": "https://sciencetostartup.com/paper/memo-memory-augmented-model-context-optimization-for-robust-multi-turn-multi-agent-llm-games", "sameAs": "https://arxiv.org/abs/2603.09022", "identifier": { "@type": "PropertyValue", "propertyID": "arXiv", "value": "2603.09022" }, "isAccessibleForFree": true, "isPartOf": { "@id": "https://sciencetostartup.com/#website" }, "datePublished": "2026-03-09T23:36:32.000Z", "additionalProperty": [ { "@type": "PropertyValue", "propertyID": "viabilityScore", "value": 3 }, { "@type": "PropertyValue", "propertyID": "researchDomain", "value": "Agents" } ] }, { "@type": "BreadcrumbList", "itemListElement": [ { "@type": "ListItem", "position": 1, "name": "Home", "item": "https://sciencetostartup.com" }, { "@type": "ListItem", "position": 2, "name": "Agents", "item": "https://sciencetostartup.com/topics" }, { "@type": "ListItem", "position": 3, "name": "MEMO: Memory-Augmented Model Context Optimization for Robust", "item": "https://sciencetostartup.com/paper/memo-memory-augmented-model-context-optimization-for-robust-multi-turn-multi-agent-llm-games" } ] } ] }

MEMO: Memory-Augmented Model Context Optimization for Robust Multi-Turn Multi-Agent LLM Games

MEMO: Memory-Augmented Model Context Optimization for Robust Multi-Turn Multi-Agent LLM Games

Claim map

Constellation map

Competitive landscape

Buzz

PDF

REFERENCES

Related Papers

Related Resources

Subscribe to the weekly brief

Build artifacts

Brief

Experiment plan

Validation checklist

Scientific founder

Translational engineer

Domain operator

GTM lead

Regulatory/clinical advisor

Timeline

Claim map

Constellation map

Competitive landscape

Buzz

PDF

REFERENCES

Related Papers

Related Resources

Subscribe to the weekly brief

Build artifacts

Brief

Experiment plan

Validation checklist

Scientific founder

Translational engineer

Domain operator

GTM lead

Regulatory/clinical advisor

Timeline