The study evaluates how to best split tasks among large‑language‑model agents for cross‑border VAT determination, comparing one broad agent to configurations ranging from one to five narrow agents. Across 4,400 runs—including token‑matched and failure‑injection scenarios—the intermediate configurations achieved the highest accuracy but did not surpass the fine‑endpoint benchmark, leaving the optimal decomposition hypothesis unconfirmed. The pilot provides a preregistered heuristic for right‑sizing decomposition, along with an oracle, dataset, and analysis pipeline.
By Pedro Santos
arXiv:2607. 17531v1 Announce Type: cross Abstract: Test-time collaboration, including self-consistency, best-of-N selection, critic models, and verifier pipelines, is often credited with broadly improving LLM reasoning, yet its gains are uneven and sometimes negative.
By Jie Hu
arXiv:2607. 17136v1 Announce Type: cross Abstract: Agentic computer-use RL is reported in single runs, and those numbers mislead.
By Barada Sahu (Cabal AI), Shivesh Pandey (Para AI)
Atlas is a framework that optimizes the deployment of compound AI workflows on heterogeneous clusters by selecting execution plans that satisfy service level objectives (SLOs). It introduces MAP, a Markovian Accuracy Predictor, which estimates configuration accuracy using local conditional accuracy transitions between adjacent workflow stages, avoiding exhaustive end‑to‑end profiling. Atlas formulates plan selection as a mixed‑integer linear program, achieving near‑oracle accuracy while reducing deployment cost by up to 42% and profiling cost by up to 2.6×.
By Milos Gravara, Andrija Stanisic, Stefan Nastic
arXiv:2606. 04161v1 Announce Type: new Abstract: Different predictors often excel on different inputs, so picking the best one per instance promises higher accuracy than committing to a single model.
By Tyler Crosse, Alan Nadelsticher Ruvalcaba, Dustin Khang LeDuc, Thomas Trask, Nicholas Lytle, David Joyner
arXiv:2606. 04402v1 Announce Type: new Abstract: Modern reasoning models can allocate different amounts of test-time computation, such as thinking tokens, model calls, or compute budget, to different tasks.
By Jingbo Wen, Liang He, Ziqi He
arXiv:2607. 17240v1 Announce Type: new Abstract: When does a committed intermediate stage in an LLM reasoning pipeline earn its cost?
By Honglin Li (ShanghaiTech University)
Agentic AI systems often approach the same task through multiple workflows that differ in reasoning strategy, verification structure, and compute cost. A natural deployment policy is to use the workfl...
arXiv:2606. 19808v1 Announce Type: new Abstract: Test-time reasoning is increasingly used as a serving-time control knob, but extra reasoning is not uniformly valuable: it can repair failed attempts, waste compute on already-correct answers, or introduce harmful answer changes.
By Sajib Acharjee Dip, Dawei Zhou, Liqing Zhang
The paper introduces the Weighted Variability Score (WVS), a metric that aggregates dispersion in runtime, optimality gap, LP iterations, and branch‑and‑bound nodes across repeated runs of MIP solvers. Experiments on 100 randomized runs of MIPLIB 2017 instances show a right‑skewed distribution, with many instances stable but some exhibiting large variability not explained by basic properties. A tree‑based predictor using static and root‑level features achieves a mean absolute error of 0.058 and Spearman rank correlation of 0.65, with root degeneracy, matrix structure, coefficient ranges, and presolve reductions identified as key drivers of high WVS.
By Mourchid Adegbindin, Brenda Dietrich, Jamol Pender, Andrea Lodi
The paper introduces a method to certify selective prediction in machine learning systems by computing the availability of safety gates through exact-binomial inversion and dynamic programming. It demonstrates that a truth-informed planner can significantly improve mean coverage over naive approaches, and that reallocating error budgets further enhances coverage across diverse applications such as LLM tool‑calling, content moderation, lesion classification, and recommendation. The study highlights the importance of planning and finite‑sample estimation in ensuring reliable, granular deployment of selective predictors.
By Parivesh Priye, Yufeng Wang, Haibin Ling, Michael Chaykowsky
The paper investigates whether providing candidate solutions during test‑time aggregation improves or harms accuracy compared to a fresh solve that does not use any candidates. Using Qwen3‑4B on AIME‑2025 and HMMT‑2025, the authors find that conditioning on multiple correct candidates boosts accuracy (+0.290), while conditioning on an all‑wrong candidate pool reduces accuracy (−0.123); the effect for a single correct candidate remains unclear. The study also explores structured interventions and placebo controls, but the underlying mechanisms of these effects are not resolved.
By Guiv Farmanfarmaian