multi-agent-llm-workflows-mixed-size-model-zoo-choosing-which-model CLM-0614In multi-agent LLM workflows with a mixed-size model zoo, choosing which model instantiates each agent step-by-step from an online estimate of task-progress gain plus remaining time and cost budget keeps task quality near the best single-model policy at lower energy than one-shot query-level routing or stall-triggered escalation.
Capability: Getting more from several models than from one
Observed on
Code generation (HumanEval Plus, MBPP), math (MATH-500) and retrieval-augmented long-form QA (ASQA), with 7-9 open-weight models per zoo and an imposed long-term energy budget; not.
Sources
- Validated by: Pass rate / citation precision, energy in Joules and wall-clock time against single-model policies, MasRouter (one-shot), CASCADIA (escalation) and an Educated Guessing baseline on four benchmarks, with a fixed long-term energy budget marking feasible methods; best or near-best among budget-satisfying methods on 3 of 4 datasets, and 3.5 points below CASCADIA on MATH-500 at lower energy.. Measured on four benchmarks with 100-200 items each (164 for HumanEval Plus), single runs reported without variance, authors' own method versus three baselines. [truncated]
Disagreeing is the most useful thing you can do here. Nothing on this page waits for a reviewer: the counts above move when evidence lands, and every source stays attached with who brought it.
Related claims
- In multi-agent LLM pipelines over reasoning and code tasks with a pool of 3B-32B open-weight backbones, conditioning each per-step routing decision on a learned filtered memory rather than the full accumulated trajectory leaves accuracy about the same while cutting inference cost by roughly 40-50%, so most of the accumulated history is cost without accuracy value.Getting more from several models than from one · reviewed by AI
- For LLM agents on dependent multi-step tasks, success decays geometrically with the number of steps under a single per-step reliability parameter that rises with model scale but stays below 1, so long horizons eventually collapse even for the strongest deployed systems — and the decay tracks step count, not context length, since bounding the context window steepens the decay rather than easing it.Completing tasks that take many steps or many hours · reviewed by AI
- For LLM agents that must compose and then execute domain tool workflows, training separate adapters for the two roles splits the gains cleanly: supervised fine-tuning of the planner on reference tool sequences mainly improves workflow-order fidelity, while node-level RL on argument-match rewards mainly improves tool-argument grounding and final-answer quality, and combining both beats either on a 9B backbone.Planning a sequence of actions before acting · reviewed by AI
- For LLM web agents on cluttered pages containing many concurrent interface components, success falls sharply as component count grows, and having the agent first emit a localized sub-goal plus a task-relevant sub-observation before choosing an action recovers much of the loss, because self-assessment of the action is more reliable in the reduced context.Using the tools it is given · reviewed by AI
- When asked to generate a full procedural execution graph for an unseen task (recipe) in a specific creator's style, few-shot LLMs assign ingredients to steps better than a statistical two-stage pipeline but predict the number of steps worse, and swapping in a larger LLM backbone does not close that step-count gap.Planning a sequence of actions before acting · reviewed by AI
Notes
Drafted from the paper by a model and filed unreviewed. Visible here so it can be read, not because anyone has vouched for it: it does not move any technique's standing and does not count toward the internal scorecard. Drafted confidence: low. Falsifier as drafted: A matched comparison where step-wise progress-guided routing fails to beat one-shot query-level routing on quality at equal or lower energy, or where its gain disappears once the progress scorer's domain-specific milestone definitions are removed.