NVIDIA paper: multi-model ensembles often underperform, single-family selection wins

omarsar0 · x · 2026-09-16

An NVIDIA team studied how to choose models for multi-agent systems, comparing eight selection strategies—by size, accuracy, answer diversity and error diversity—across routing, majority vote and LLM-as-judge setups on hard science benchmarks.

Key findings:

Takeaway: before adding another model to a router or ensemble, measure what it adds.

Original post →

More from Research

Research channel →