Multi-Agent Reasoning Improves Compute Efficiency: Pareto-Optimal Test-Time Scaling

• URL: http://arxiv.org/abs/2605.01566v1
• Date: Sat, 02 May 2026 18:31:02 GMT
• Title: Multi-Agent Reasoning Improves Compute Efficiency: Pareto-Optimal Test-Time Scaling
• Abstract summary: computational efficiency is key for real-world applications with resource constraints.<n>We provide a systematic analysis of the inference scaling strategies self-consistency, self-refinement, multi-agent debate, and mixture-of-agents.<n>We identify a simple multi-agent design guideline: mixture-of-agents is most efficient when the number of parallel generations exceeds the number of sequential aggregations.
• Score: 11.219930588268433
• License: http://creativecommons.org/licenses/by-sa/4.0/
• Abstract: Advances in inference methods have enabled language models to improve their predictions without additional training. These methods often prioritize raw performance over cost-effective compute usage. However, computational efficiency is key for real-world applications with resource constraints. We provide a systematic analysis of the inference scaling strategies self-consistency, self-refinement, multi-agent debate, and mixture-of-agents, to study their computational performance tradeoffs. We evaluate methods on two reasoning benchmarks (MMLU-Pro, BBH) and include extensive parameter configurations (e.g., scaling the number of parallel predictions, agents, and debate rounds) across different model sizes. Across 34 configurations and over 100 evaluations, we compute the Pareto-optimal front to select methods that achieve the best accuracy with the lowest computational budget. Notably, inference scaling improves accuracy by up to +7.1% points over chain-of-thought at the highest evaluated budgets (20x the CoT compute budget) on MMLU-Pro. With an equal computing budget, debate and mixture-of-agents outperform self-consistency by 1.3% and 2.7% points, respectively. While self-consistency saturates earlier, multi-agent gains persist, particularly on more complicated tasks. We identify a simple multi-agent design guideline: mixture-of-agents is most efficient when the number of parallel generations exceeds the number of sequential aggregations.

Related papers

• How Inference Compute Shapes Frontier LLM Evaluation [2.7633794124150426]
  We evaluate up to 12 frontier language models on seven challenging benchmarks spanning software engineering, mathematics, medicine, and cybersecurity.<n>We find that newer models reach higher performance at large budgets, where they unlock harder tasks and solve them more reliably.<n>We argue that evaluations should report capability as a function of inference-time compute, specify protocol choices explicitly, and compare model generations over a large shared compute range.
  arXiv  Detail & Related papers  (2026-06-16T13:40:53Z)
• InferenceBench: A Benchmark for Open-Ended LLM Inference Optimization by AI Agents [15.892896819368751]
  We introduce InferenceBench, where an agent must deploy an OpenAI-compatible inference server and optimize the speed of LLM inference.<n>Three optimization scenarios isolate distinct bottlenecks of inference.<n>Agents reliably improve over a naive PyTorch baseline.<n>Overall, InferenceBench reflects the ability of agents to operate in an open-ended AI engineering setting.
  arXiv  Detail & Related papers  (2026-05-20T15:55:21Z)
• Single-Agent LLMs Outperform Multi-Agent Systems on Multi-Hop Reasoning Under Equal Thinking Token Budgets [14.028807574426748]
  We present an information-theoretic argument, grounded in the Data Processing Inequality.<n>We find that SAS consistently match or outperform MAS on multi-hop reasoning tasks when reasoning tokens are held constant.<n>Overall, our results suggest that, for multi-hop reasoning tasks, many reported advantages of multi-agent systems are better explained by unaccounted computation and context effects.
  arXiv  Detail & Related papers  (2026-04-02T18:47:48Z)
• $V_1$: Unifying Generation and Self-Verification for Parallel Reasoners [69.66089681814013]
  $V_$ is a framework that unifies generation and verification through efficient pairwise ranking.<n>$V_$-Infer improves Pass@1 by up to $10%$ over pointwise verification.<n>$V_$-PairRL achieves $7$--$9%$ test-time scaling gains over standard RL and pointwise joint training.
  arXiv  Detail & Related papers  (2026-03-04T17:22:16Z)
• Para-B&B: Load-Balanced Deterministic Parallelization of Solving MIP [50.917107318582715]
  Mixed-integer programming (MIP) extends linear programming by incorporating both continuous and integer decision variables.<n>This paper presents the first fully open-source implementation of deterministic parallel branch-and-bound for HiGHS, a high-performance MIP solver.<n>Our approach introduces a novel data-parallel architecture ensuring strict determinism by replicating complete solver state across worker threads.
  arXiv  Detail & Related papers  (2026-02-10T14:17:53Z)
• Towards a Science of Scaling Agent Systems [79.64446272302287]
  We formalize a definition for agent evaluation and characterize scaling laws as the interplay between agent quantity, coordination structure, modelic, and task properties.<n>We derive a predictive model using coordination metrics, that cross-validated R2=0, enabling prediction on unseen task domains.<n>We identify three effects: (1) a tool-coordination trade-off: under fixed computational budgets, tool-heavy tasks suffer disproportionately from multi-agent overhead, and (2) a capability saturation: coordination yields diminishing or negative returns once single-agent baselines exceed 45%.
  arXiv  Detail & Related papers  (2025-12-09T06:52:21Z)
• DeepPrune: Parallel Scaling without Inter-trace Redundancy [53.62015294143274]
  Over 80% of parallel reasoning traces yield identical final answers, representing substantial wasted computation.<n>We propose DeepPrune, a novel framework that enables efficient parallel scaling through dynamic pruning.<n>Our work establishes a new standard for efficient parallel reasoning, making high-performance reasoning more efficient.
  arXiv  Detail & Related papers  (2025-10-09T17:24:54Z)
• Optimizing Anytime Reasoning via Budget Relative Policy Optimization [70.32755424260336]
  We present a novel framework, AnytimeReasoner, to optimize anytime reasoning performance.<n>We truncate the complete thinking process to fit within sampled token budgets from a prior distribution.<n>We then optimize the thinking and summary policies in a decoupled manner to maximize the cumulative reward.
  arXiv  Detail & Related papers  (2025-05-19T17:58:44Z)
• Learning Adaptive Parallel Reasoning with Language Models [70.1745752819628]
  We propose Adaptive Parallel Reasoning (APR), a novel reasoning framework that enables language models to orchestrate both serialized and parallel computations end-to-end.<n> APR generalizes existing reasoning methods by enabling adaptive multi-threaded inference using spawn() and join() operations.<n>A key innovation is our end-to-end reinforcement learning strategy, optimizing both parent and child inference threads to enhance task success rate without requiring predefined reasoning structures.
  arXiv  Detail & Related papers  (2025-04-21T22:29:02Z)
• Entropy-Gated Branching for Efficient Test-Time Reasoning [21.810952984561116]
  Test-time compute methods can significantly improve the reasoning capabilities and problem-solving accuracy of large language models (LLMs)<n>We propose Entropy-Gated Branching (EGB), which branches only at high-uncertainty steps and prunes expansions with a lightweight verifier.<n>On mathematical and financial reasoning benchmarks, EGB improves accuracy by 22.6% over standard inference while operating 31%-75% faster across math benchmarks than test-time beam search with higher performance.
  arXiv  Detail & Related papers  (2025-03-27T20:18:22Z)
• Do NOT Think That Much for 2+3=? On the Overthinking of o1-Like LLMs [76.43407125275202]
  o1-like models can emulate human-like long-time thinking during inference.<n>This paper presents the first comprehensive study on the prevalent issue of overthinking in these models.<n>We propose strategies to mitigate overthinking, streamlining reasoning processes without compromising accuracy.
  arXiv  Detail & Related papers  (2024-12-30T18:55:12Z)
• Inference Scaling vs Reasoning: An Empirical Analysis of Compute-Optimal LLM Problem-Solving [0.0]
  Recent advances in large language models (LLMs) have predominantly focused on maximizing accuracy and reasoning capabilities.<n>This paper investigates the potential synergy between reasoning enhancement and computational efficiency by analyzing the integration of two contrasting approaches.
  arXiv  Detail & Related papers  (2024-12-20T08:42:45Z)
• Best Arm Identification in Stochastic Bandits: Beyond $\beta-$optimality [31.359578768463752]
  This paper investigates a hitherto unaddressed aspect of best arm identification (BAI) in multi-armed bandits in the fixed-confidence setting. Two key metrics for assessing bandit algorithms are computational efficiency and performance optimality. This paper introduces a framework and an algorithm for BAI that achieves optimal performance with a computationally efficient set of decision rules.
  arXiv  Detail & Related papers  (2023-01-10T05:02:49Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.