# When Does LLM Orchestration Pay Off? A Controlled Evaluation of Accuracy, Cost, and Task Difficulty

arXiv: https://arxiv.org/abs/2608.00685 · HTML: https://mlsanity.com/paper/2608.00685/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 26 of its 30 references that are in the map.
- First posted on arXiv 1 Aug 2026.

## Builds on (most cited first, 26 of 26)

- [Chain-of-Thought Prompting Elicits Reasoning in Large Language Models](https://mlsanity.com/paper/2201.11903/) (2022)
- [Large Language Models are Zero-Shot Reasoners](https://mlsanity.com/paper/2205.11916/) (2022)
- [Self-Refine: Iterative Refinement with Self-Feedback](https://mlsanity.com/paper/2303.17651/) (2023)
- [Improving Factuality and Reasoning in Language Models through Multiagent Debate](https://mlsanity.com/paper/2305.14325/) (2023)
- [Large Language Models Cannot Self-Correct Reasoning Yet](https://mlsanity.com/paper/2310.01798/) (2023)
- [RouteLLM: Learning to Route LLMs with Preference Data](https://mlsanity.com/paper/2406.18665/) (2024)
- [Why Do Multi-Agent LLM Systems Fail?](https://mlsanity.com/paper/2503.13657/) (2025)
- [BigCodeBench: Benchmarking Code Generation with Diverse Function Calls and Complex Instructions](https://mlsanity.com/paper/2406.15877/) (2024)
- [GEPA: Reflective Prompt Evolution Can Outperform Reinforcement Learning](https://mlsanity.com/paper/2507.19457/) (2025)
- [AFlow: Automating Agentic Workflow Generation](https://mlsanity.com/paper/2410.10762/) (2024)
- [Towards Reasoning Era: A Survey of Long Chain-of-Thought for Reasoning Large Language Models](https://mlsanity.com/paper/2503.09567/) (2025)
- [MathArena: Evaluating LLMs on Uncontaminated Math Competitions](https://mlsanity.com/paper/2505.23281/) (2025)
- [A Comprehensive Survey of Self-Evolving AI Agents: A New Paradigm Bridging Foundation Models and Lifelong Agentic Systems](https://mlsanity.com/paper/2508.07407/) (2025)
- [More Agents Is All You Need](https://mlsanity.com/paper/2402.05120/) (2024)
- [Multi-agent Architecture Search via Agentic Supernet](https://mlsanity.com/paper/2502.04180/) (2025)
- [Should we be going MAD? A Look at Multi-Agent Debate Strategies for LLMs](https://mlsanity.com/paper/2311.17371/) (2023)
- [A Survey on the Optimization of Large Language Model-based Agents](https://mlsanity.com/paper/2503.12434/) (2025)
- [Single-Agent LLMs Outperform Multi-Agent Systems on Multi-Hop Reasoning Under Equal Thinking Token Budgets](https://mlsanity.com/paper/2604.02460/) (2026)
- [Difficulty-Aware Agentic Orchestration for Query-Specific Multi-Agent Workflows](https://mlsanity.com/paper/2509.11079/) (2025)
- [Easy2Hard-Bench: Standardized Difficulty Labels for Profiling LLM Performance and Generalization](https://mlsanity.com/paper/2409.18433/) (2024)
- [OckBench: Measuring the Efficiency of LLM Reasoning](https://mlsanity.com/paper/2511.05722/) (2025)
- [Scaling Graph Chain-of-Thought Reasoning: A Multi-Agent Framework with Efficient LLM Serving](https://mlsanity.com/paper/2511.01633/) (2025)
- [MAS-PromptBench: When Does Prompt Optimization Improve Multi-Agent LLM Systems?](https://mlsanity.com/paper/2606.23664/) (2026)
- [The Agentic Researcher: A Practical Guide to AI-Assisted Research in Mathematics and Machine Learning](https://mlsanity.com/paper/2603.15914/) (2026)
- [ComplexityNet: Increasing LLM Inference Efficiency by Learning Task Complexity](https://mlsanity.com/paper/2312.11511/) (2023)
- [AdaptiveLLM: A Framework for Selecting Optimal Cost-Efficient LLM for Code-Generation Based on CoT Length](https://mlsanity.com/paper/2506.10525/) (2025)

## Abstract (arXiv)

LLM orchestration is often assumed to improve reasoning by allocating additional inference-time computation, yet its gains may not justify its cost. Existing comparisons also frequently overlook differences in optimization effort, making it difficult to isolate the value of orchestration itself. We conduct a controlled evaluation of Self-Refine, Best-of-N, and Debate against task-only and chain-of-thought (CoT) single-call baselines across five LLM backbones and three domains: competitive programming, chess puzzles, and mathematics. For comparability, we optimize each method with GEPA under the same optimization budget and evaluate all methods on the same difficulty-stratified benchmark items. Orchestration yields moderate but benchmark-dependent gains: averaged across backbones within each benchmark, the largest improvement is 4.6 percentage points over optimized CoT inference and 4.5 points over task-only inference, while requiring approximately 2 to 4 times the mean total tokens of task-only inference. Human-derived difficulty is associated with lower absolute accuracy in all three benchmarks, but within-benchmark analyses do not indicate that orchestration effects increase with task difficulty. By contrast, exploratory mixed-effects analyses reveal strong interactions between orchestration method and backbone model across all three benchmarks, showing that orchestration effectiveness depends substantially on the underlying model. Our results suggest that orchestration decisions should be model-specific and account for whether moderate accuracy gains justify the additional inference cost. More broadly, evaluations of LLM orchestrations should control optimization effort and report model-specific accuracy--cost trade-offs rather than treating additional inference-time structure as uniformly beneficial.