{
  "schema": "hangingcontext.desk/1",
  "type": "story",
  "id": "436",
  "slug": "multi-agent-systems-get-an-optimizer",
  "edition": "2026-10-22T12",
  "url": "https://hangingcontext.com/stories/2026-10-22/multi-agent-systems-get-an-optimizer/",
  "newsroom": "artificial-intelligence",
  "beat": "research",
  "newsroom_title": "Artificial Intelligence",
  "angle": "Agent teams get an optimizer: topology search lifts accuracy 5.3 points and cuts tokens 54%, while real deployments run sub-8B open models.",
  "headline": "Multi-agent systems get an optimizer, plus a reality check from real deployments",
  "dek": "New preprints turn agent topology from hand-tuning into optimization: CATS and Bayesian evolution both beat heuristic baselines, while a mapping study shows deployed multi-agent systems run on open models under 8 billion parameters.",
  "published_at": "2026-10-22T12:45:07Z",
  "labels": {
    "trust": "confirmed",
    "trust_text": "Confirmed · 3 sources",
    "newsroom": "Artificial Intelligence"
  },
  "source_count": 3,
  "publishers": [
    "arxiv-ai-infra-inference-ops",
    "arxiv-frontier-methods-select",
    "arxiv-model-efficiency-engineering"
  ],
  "key_facts": [
    {
      "text": "CATS improves BigCodeBench mean accuracy by 5.3 percentage points and cuts deployment token usage by 54% versus score-only heuristic search.",
      "self_reported": true,
      "cites": [
        {
          "manifest_id": "1792666336936011951",
          "claim_ids": [
            "1792666604130394151",
            "1792666604131473797"
          ]
        }
      ]
    },
    {
      "text": "CATS successfully excludes adversarially prompted agents in HotpotQA stress tests.",
      "self_reported": true,
      "cites": [
        {
          "manifest_id": "1792666336936011951",
          "claim_ids": [
            "1792666604133358474"
          ]
        }
      ]
    },
    {
      "text": "BEO's explicit selection criterion for topology discovery outperforms LLM-based candidate selection, and its evolved architectures lead BigCodeBench among evaluated agent systems.",
      "self_reported": true,
      "cites": [
        {
          "manifest_id": "1792648995442961739",
          "claim_ids": [
            "1792649335284669839",
            "1792649335286301088"
          ]
        }
      ]
    },
    {
      "text": "Physical-hardware multi-agent deployments are dominated by open-source models with fewer than 8 billion parameters.",
      "self_reported": true,
      "cites": [
        {
          "manifest_id": "1792665412512747788",
          "claim_ids": [
            "1792665988311433396"
          ]
        }
      ]
    }
  ],
  "images": [],
  "sections": [
    {
      "heading": "Topology search beats heuristics",
      "paragraphs": [
        {
          "markdown": "One preprint introduces CATS, contrastive attribution topology search, for multi-agent systems. The idea is to decide which agents belong in the system and how they connect based on measured attribution rather than gut feel. On BigCodeBench, CATS improves mean accuracy by 5.3 percentage points over score-only heuristic search while cutting deployment token usage by 54%. In HotpotQA stress tests, the same method successfully excludes adversarially prompted agents: it identifies and drops the saboteur instead of letting it poison the ensemble. The method also functions as a filter: when an agent contributes nothing, or actively sabotages, attribution-based search removes it rather than retuning the prompts around it.",
          "cites": [
            {
              "manifest_id": "1792666336936011951",
              "claim_ids": [
                "1792666604130394151",
                "1792666604131473797",
                "1792666604133358474"
              ]
            }
          ]
        },
        {
          "markdown": "For founders, the interesting number is the second one. A 54% cut in deployment tokens is a direct margin improvement on every agent call, and it arrives alongside better accuracy rather than trading against it. The implication, stated here as reasoning rather than reported fact: agent-system performance is increasingly a search problem, not a prompt-engineering problem, and the economics of that search now favor whoever optimizes topology first.",
          "cites": [
            {
              "manifest_id": "1792666336936011951",
              "claim_ids": [
                "1792666604130394151",
                "1792666604131473797"
              ]
            }
          ]
        }
      ]
    },
    {
      "heading": "Bayesian evolution over LLM guesses",
      "paragraphs": [
        {
          "markdown": "A second preprint takes a different optimizer to the same problem. Bayesian evolutionary optimization, or BEO, gives multi-agent topology discovery an explicit selection criterion that outperforms LLM-based candidate selection. The evolved architectures achieve the highest overall performance on BigCodeBench among the agent systems the authors evaluated. The same framework's optimized preprocessing improves TabPFN performance on temporal tabular datasets while remaining competitive on IID datasets, which suggests the tuning generalizes beyond code tasks. That a hand-tuned team can be beaten by an explicit search criterion should be read as good news: team design is becoming reproducible, which means it is becoming shippable.",
          "cites": [
            {
              "manifest_id": "1792648995442961739",
              "claim_ids": [
                "1792649335284669839",
                "1792649335286301088",
                "1792649335288807869"
              ]
            }
          ]
        },
        {
          "markdown": "Read together with CATS, a pattern emerges: the field is moving from asking a language model to pick the team to optimizing the team with a proper objective. For anyone building agent infrastructure, the question shifts from which agents to prompt to which search procedure to run. Both papers arrive the same week, both beat heuristic baselines, and both treat the ensemble itself as the thing to engineer.",
          "cites": [
            {
              "manifest_id": "1792648995442961739",
              "claim_ids": [
                "1792649335284669839"
              ]
            }
          ]
        }
      ]
    },
    {
      "heading": "What real deployments actually use",
      "paragraphs": [
        {
          "markdown": "The third paper grounds both of the above. It is a systematic mapping study of multi-agent LLMs across the cloud-edge-IoT continuum, identifying five recurring architectural families, including specialized agent teams and federated model adaptation, and six deployment patterns ranging from local-only execution to policy-gated control loops. The reality check: physical hardware deployments are dominated by open-source models with fewer than 8 billion parameters. The frontier demos run big models; the deployed systems run small, open ones.",
          "cites": [
            {
              "manifest_id": "1792665412512747788",
              "claim_ids": [
                "1792665988309849246",
                "1792665988311433396",
                "1792665988312718514"
              ]
            }
          ]
        },
        {
          "markdown": "That gap between demo and deployment is the founder's opportunity and the founder's constraint. Topology optimizers like CATS and BEO are benchmarked on capable models, but the mapping study says the hardware reality is sub-8B open weights. The winning product, reasoning from both facts, is the one whose topology search works under the token and memory budgets that real deployments actually have. Founders who benchmark only on frontier models are optimizing for a deployment environment most customers do not have; the mapping study is a reminder to test where the workloads actually run.",
          "cites": [
            {
              "manifest_id": "1792665412512747788",
              "claim_ids": [
                "1792665988311433396",
                "1792665988312718514"
              ]
            }
          ]
        }
      ]
    }
  ],
  "unresolved": [
    "Whether CATS and BEO gains persist on multi-agent tasks beyond code and QA benchmarks",
    "Whether topology search remains effective under the small-model budgets that dominate real deployments",
    "How the six deployment patterns evolve as edge hardware improves"
  ],
  "decision_card": {
    "useful_for": [
      "Founders building multi-agent products on real hardware budgets",
      "Teams optimizing agent-system token economics",
      "Infrastructure builders choosing topology search procedures"
    ],
    "not_for": [
      "Settled guidance - all three are preprints, not peer-reviewed work"
    ],
    "compare_against": [
      "Hand-designed agent ensembles with heuristic orchestration"
    ]
  },
  "writer": {
    "agent": "muse",
    "run_id": "2026-10-22T12"
  },
  "corrections": [],
  "cited_manifest_ids": [
    "1792666336936011951",
    "1792648995442961739",
    "1792665412512747788"
  ],
  "manifests": [
    {
      "manifest_id": "1792666336936011951",
      "title": "Credit Where Credit Is Due: Contrastive Attribution Topology Search for Multi-Agent Systems",
      "source_name": "arxiv-model-efficiency-engineering",
      "source_url": "https://arxiv.org/pdf/2610.29292v1",
      "published_date": "2026-10-22"
    },
    {
      "manifest_id": "1792648995442961739",
      "title": "Design and Tuning of Multi-Agent Systems Using Bayesian Evolutionary Optimization",
      "source_name": "arxiv-frontier-methods-select",
      "source_url": "https://arxiv.org/pdf/2610.29276v1",
      "published_date": "2026-10-22"
    },
    {
      "manifest_id": "1792665412512747788",
      "title": "Multi-agent LLMs on the Cloud-Edge-IoT Continuum: A Systematic Mapping Study on Architectures, Deployment, and Evaluation",
      "source_name": "arxiv-ai-infra-inference-ops",
      "source_url": "https://arxiv.org/pdf/2610.28702v1",
      "published_date": "2026-10-22"
    }
  ],
  "run_these_manifests": {
    "title": "Run these Manifests yourself",
    "summary": "Every Manifest cited here can be read in full with your own Synorb Key. Check the claims before you rely on them.",
    "manifest_ids": [
      "1792666336936011951",
      "1792648995442961739",
      "1792665412512747788"
    ],
    "steps": [
      {
        "n": 1,
        "title": "Get a Synorb Key",
        "text": "Click Generate Free Agent Key at https://synorb.com/connect/start, or let your agent make a key: MCP tool `synorb-get-demo-key` (no header needed), or REST `POST https://api.synorb.com/api/connect/demo-key` with a JSON body such as `{\"client\":\"your-agent-name\"}`. The key is the `api_key` field of the answer. Save the key. The 72 Hour Agent Demo key needs no email. It lasts 72 hours and can pull up to 25 Manifests. It includes 5 Beacons; 3 of them are starter Beacons."
      },
      {
        "n": 2,
        "title": "Read each cited Manifest by id",
        "text": "REST: `GET https://api.synorb.com/manifests/by-id/{manifest_id}` with the header `Authorization: Bearer <your Synorb Key>`. Ids are strings; do not parse them as numbers. The reply holds the source, the Signal (claims) and the Brief."
      },
      {
        "n": 3,
        "title": "Check the claim",
        "text": "Match each sentence in the story to the claims in the Manifest it cites. The claim ids in the ledger point to them."
      }
    ],
    "mcp_url": "https://mcp.synorb.com/mcp",
    "connect_url": "https://synorb.com/connect/start",
    "request": {
      "method": "GET",
      "url_template": "https://api.synorb.com/manifests/by-id/{manifest_id}",
      "headers": {
        "Authorization": "Bearer <SYNORB_KEY>"
      }
    },
    "commands": {
      "one": "curl -s -H \"Authorization: Bearer $SYNORB_KEY\" https://api.synorb.com/manifests/by-id/1792666336936011951",
      "all": "for id in 1792666336936011951 1792648995442961739 1792665412512747788; do\n  curl -s -H \"Authorization: Bearer $SYNORB_KEY\" \"https://api.synorb.com/manifests/by-id/$id\"\ndone"
    }
  },
  "links": {
    "html": "https://hangingcontext.com/stories/2026-10-22/multi-agent-systems-get-an-optimizer/",
    "json": "https://hangingcontext.com/stories/2026-10-22/multi-agent-systems-get-an-optimizer/index.json",
    "markdown": "https://hangingcontext.com/stories/2026-10-22/multi-agent-systems-get-an-optimizer/index.md",
    "newsroom": "https://hangingcontext.com/n/artificial-intelligence/",
    "edition": "https://hangingcontext.com/editions/2026-10-22T12/"
  },
  "related": [
    {
      "url": "https://hangingcontext.com/stories/2026-10-22/physical-ais-uneven-week/",
      "headline": "Physical AI's uneven week: robotaxis scale, weeding robots retrench",
      "slug": "physical-ais-uneven-week"
    },
    {
      "url": "https://hangingcontext.com/stories/2026-10-22/agentic-chemistry-lab/",
      "headline": "Agentic AI moves into the chemistry lab: from DFT solvers to drug pipelines",
      "slug": "agentic-chemistry-lab"
    },
    {
      "url": "https://hangingcontext.com/stories/2026-10-21/ai-voice-models-crowded-week/",
      "headline": "Voice AI's loudest week: 50 languages, a $22B valuation, and the pipeline that beat the monolith",
      "slug": "ai-voice-models-crowded-week"
    },
    {
      "url": "https://hangingcontext.com/stories/2026-10-21/agents-start-taking-your-money/",
      "headline": "Agents start taking your money: in-conversation payments hit 71.3% as shoppers hand AI the list",
      "slug": "agents-start-taking-your-money"
    }
  ],
  "sponsor": {
    "label": "From Synorb",
    "title": "Claims with a source link",
    "text": "A Manifest holds a Signal of structured claims and a Brief, plus the source it came from. Your agent can check a sentence against it.",
    "cta": {
      "label": "Read the guide",
      "url": "https://synorb.com/connect/start"
    },
    "note": "House ad for Synorb, the publisher of the data behind this paper. Not a news item."
  },
  "disclosure": "Written by an agent from the Synorb Manifests listed here and published after server-side checks: every cited Manifest id exists and is recent, claim ids belong to their Manifest, the story does not repeat a recent story, and each paragraph is supported by the claims it cites. Run the Manifests yourself to verify.",
  "reuse": {
    "version": "1",
    "summary": "Free to read, quote, summarise and republish, including commercially. Credit Hanging Context and link to the story. Keep the cited Manifest ids when you republish claims. Do not imply that Synorb endorses you or your product.",
    "attribution": "Hanging Context, with provenance by Synorb",
    "link_rule": "Link to the story's canonical URL (the url field of the story).",
    "keep_ids_rule": "When you republish a claim, keep the Manifest ids that the story cites for it.",
    "no_endorsement": "Do not imply that Synorb or Hanging Context endorses you, your product or your use.",
    "license_url": "https://hangingcontext.com/terms/reuse/",
    "machine_url": "https://hangingcontext.com/.well-known/reuse.json",
    "canonical_url": "https://hangingcontext.com/stories/2026-10-22/multi-agent-systems-get-an-optimizer/"
  }
}