A preprint holds model, tasks and tools fixed and changes only the agent architecture; multi-agent setups use 1.4 to 4.7 times the tokens of a single agent, mostly with no significant difference in success rate.

Concept illustration of an open three-walled display box on a black base lettered ARXIV in white. On the back wall, a sign reading SINGLE AGENT sits above one laptop whose screen shows a speech bubble, with a short grey stack of flat slabs to its right, and a sign reading MULTI-AGENT sits above three laptops side by side, each screen showing a speech bubble; a tray of file folders sits in front of each group, and at the far right a much taller red stack of flat slabs rises above the walls. Printed on the image: 1.4–4.7x Multi-agent token use vs a single agent, same model and tasks, self-reported by preprint authors.

According to the paper's abstract, the only setup that significantly beat a single agent was a redesigned multi-agent implementation, a result the authors report and no one else has repeated. That setup's task success rate was 75.6%, against 67.4% for a single agent. The authors ran eight configurations of four designs on 13 enterprise standard operating procedures. Coordinated configurations had 2.6 to 5.8 times the latency of a single agent. The setup with a high-quality skill library reached 73.1%, and the one with a poor skill library fell below 67.4%. The abstract doesn't say which model was used (arXiv 2611.12953, 2026-11-07). How we checked →

Source and checks

This item ran in the brief of Nov 11, 2026, alongside 15 other items. How we checked it, the quote checks and what was still missing that day are on that issue's How we checked page.

Cite this item

SecondSource (2026). "A preprint holds model, tasks and tools fixed and changes only the agent architecture; multi-agent setups use 1.4 to 4.7 times the tokens of a single agent, mostly with no significant difference in success rate.." https://secondsource.io/en/research/2026-11-11-2/. Accessed access date.