# Test-Time Compute Scales Better with Parallel Agent Teams

DevFeed: [Test-Time Compute Scales Better with Parallel Agent Teams](<https://devfeed.tech/articles/test-time-compute-scales-better-with-parallel-agent-teams-124726.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.09798>)

Author: Junlin Wang, Haozhe Chen, Ang Li, Rahul Chalamala, Zhuang Liu

Published: 2026-11-09T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [llm-reasoning](<https://devfeed.tech/topics/llm-reasoning.md>), [Multi Agent Systems](<https://devfeed.tech/topics/multi-agent-systems.md>), [Parallel reasoning](<https://devfeed.tech/topics/parallel-reasoning.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agent-teams](<https://devfeed.tech/tags/agent-teams.md>), [agents](<https://devfeed.tech/tags/agents.md>), [best-of](<https://devfeed.tech/tags/best-of.md>), [budget](<https://devfeed.tech/tags/budget.md>), [compute](<https://devfeed.tech/tags/compute.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [development](<https://devfeed.tech/tags/development.md>), [exploit-development](<https://devfeed.tech/tags/exploit-development.md>), [formal](<https://devfeed.tech/tags/formal.md>), [frontier-model](<https://devfeed.tech/tags/frontier-model.md>), [gpu](<https://devfeed.tech/tags/gpu.md>), [horizon](<https://devfeed.tech/tags/horizon.md>), [inference](<https://devfeed.tech/tags/inference.md>), [kernels](<https://devfeed.tech/tags/kernels.md>), [model](<https://devfeed.tech/tags/model.md>), [parallel](<https://devfeed.tech/tags/parallel.md>), [performance](<https://devfeed.tech/tags/performance.md>), [teams](<https://devfeed.tech/tags/teams.md>), [test](<https://devfeed.tech/tags/test.md>), [time](<https://devfeed.tech/tags/time.md>)

## AI overview

Parallel agent teams--many small collaborating teams sampled independently, with the best team's result kept--outperformed 100 independent agents on 10 of 12 long-horizon tasks using a 27B model and 100-agent budget. On the four tasks tested against a single 100-agent team, they won all four; at equal dollar cost, they beat a frontier model on half the tasks. The authors' statistical model attributes the gains to balancing the higher mean performance of collaboration with the benefits of sampling variance, which many independent agents or one large team may fail to capture. The abstract reports results across tasks including GPU kernels, formal proofs, and exploit development.

## Source excerpt

arXiv:2611.09798v1 Announce Type: new Abstract: Test-time scaling improves the performance of a fixed model by spending more computation at inference. There are two common ways to spend it on agents when a verifier can score every attempt. Best-of-N sampling runs many independent agents and keeps the best result but saturates quickly. One large team of collaborating agents can score higher but stops improving as it grows. Can we get the best of both? Our method, parallel agent teams, does so by sampling many small teams in parallel and keeping the best team's result. We test it on 12 long-horizon tasks from GPU kernels and formal proofs to exploit development. With a 27B model and a budget of 100 agents, parallel teams of ten beat parallel agents on 10 of 12 tasks and beat one team of 100 on all four tasks where we ran such large teams. At equal dollar cost the same parallel teams beat a frontier model on half the tasks. A simple statistical model shows that individual agents lose because their many samples have a low mean and that one large team loses because a single sample gets no gain from variance. Sampling many small teams in parallel balances a high mean with the gain from variance. Parallel agent teams thus offer a more effective way to scale test-time compute.