# NVIDIA 论文：多智能体系统如何选模型池

- 来源：elvis (@omarsar0)
- 发布时间：2026-09-16 22:49
- AIHOT 分数：50
- AIHOT 链接：https://aihot.news/items/cmu48ozgg0ducro4wq29kat7t
- 原文链接：https://x.com/omarsar0/status/2100235516918849661

## AI 摘要

NVIDIA 论文对比了 8 种模型选择策略，在路由、多数投票和 LLM-as-judge 三种设置下测试模型池对多智能体系统的影响。结果显示，混用不同开源模型虽能提高理论最优准确率，但实际表现常低于池中单一最佳模型；用同一模型的多个副本反而更好。多数投票将 HLE 准确率从 29.4% 提升至 32.2%，而几乎每个混合模型组合都出现下降；从单一模型家族中选候选模型带来的提升最大。

## 正文

Banger paper from NVIDIA.

It's on the topic of choosing which models go into a multi-agent system.

The team compared eight selection strategies, based on size, accuracy, answer diversity and error diversity, across routing, majority vote and LLM-as-judge setups on hard science benchmarks.

Larger pools of different open models raised the theoretical best-case accuracy. Achieved accuracy often fell below the single best model in the pool.

Using several copies of one model worked better.

Majority vote over the best single model raised HLE accuracy from 29.4% to 32.2%, while nearly every mixed-model group declined.

Choosing candidates from a single model family gave the largest improvement over a standalone model of all eight strategies.

Before adding another model to a router or ensemble, measure what it adds.

Paper: https://arxiv.org/abs/2609.17306

Chat with Paper: https://academy.dair.ai/papers/mo-models-mo-problems-how-to-best-select-model-pools-when-designing-multi-agent-2609.17306