Vals AI 研究：AI 智能体团队成本高但质量提升有限
AI agent teams waste massive tokens for barely measurable quality gains, research finds
评测公司 Vals AI 在 Vibe Code Bench 上测试 GPT-6 Sol 和 Claude Opus 5.5，分别以单智能体和智能体团队形式运行，并设置中等与最高两档推理强度。团队方案成本是单智能体的 1.8 至 5.1 倍，四次对比中只有 GPT-6 Sol 在中等推理强度下出现统计显著提升，高出 7.3 分，最高推理强度下两种模型都无明显优势。Anthropic 在 Opus 5.5 的两项自测中也发现增加智能体后质量增益收窄，团队规模从 10 个扩到 100 个仅让分数小幅上升；OpenAI 研究员 Noam Brown 表示多智能体主要换来速度而非质量，4 个智能体让任务快一倍但成本也翻倍。
推荐理由
AI 评分 66
Vals AI 的对比显示多智能体团队成本高出 1.8 至 5.1 倍，却只在一种配置下带来统计显著提升，可据此判断团队式 Agent 的适用边界。
主题
正文
当前显示原文，中文全文译文尚未就绪。
AI agent teams deliver almost no better results than single agents, research finds.
Evals company Vals AI tested GPT-6 Sol and Claude Opus 5.5 on the "Vibe Code Bench," both solo and as teams, at two reasoning levels: medium and maximum reasoning effort. The teams cost between 1.8x and 5.1x more than single agents.
Out of four comparisons between teams and solo agents, only one showed a statistically significant improvement: GPT-6 Sol at medium reasoning, where the team scored 7.3 points higher. At maximum reasoning, the team setup gave neither Sol nor Opus 5.5 any real advantage. The results suggest that the extra cost of agent teams isn't worth it in most cases, especially when models are already running at full compute.
Vals AI's benchmark shows cost per app versus score on Vibe Code Bench. Arrows point from each model's single agent to its team at the same reasoning effort. Teams cost far more but barely improve scores. | Image: Vals AI[ Anthropic saw quality gains shrink as it added more agents in two of its own tests with Opus 5.5 . Larger teams reached a given performance level faster, but going from ten to 100 agents only nudged scores up slightly after 24 hours. In separate ProgramBench tests, speed gains came with higher token usage.
Fable 5.1 showed stronger quality gains on the Lean theorem proving task above ten agents, but still scored below Opus 5.5 across all tests. On the knowledge base task, Fable's score actually dipped slightly when scaling from 30 to 100 agents.
More agents buy speed but not better output
OpenAI researcher Noam Brown confirmed in the Dwarkesh Podcast that multi-agent systems mainly buy speed, not better quality. Four agents solved tasks twice as fast but also cost twice as much. At 16 agents, the pattern held but grew slightly less efficient.
The effect depends heavily on the task. Web research and math parallelize well, but writing a novel doesn't, he said. Throwing 10,000 agents at a novel would be just as pointless as throwing 10,000 people at it. Brown acknowledged that scaling to very large numbers of agents remains largely unexplored because the costs are simply too high.
OpenAI developer Eric Provencher recently warned against using agent swarms for exactly this reason. They're most likely wasted money, he argued, because coordination between agents breaks down. He called it the coordination tax.