Design · Orchestration
Mixture-of-Agents (MoA)
Layers of LLMs where each layer's agents take all prior-layer responses as auxiliary context, then a final aggregator synthesises. Quality compounds across layers.
Verified against arxiv.org
How it works
Layers of LLMs where each layer's agents take all prior-layer responses as auxiliary context, then a final aggregator synthesises. Quality compounds across layers.
At a glance
TechniqueEnsemble
Cost leverHigher, runs many models per query
Latency leverHigher, sequential layers
Quality leverOpen-weight ensemble beats GPT-4o on AlpacaEval 2.0
ExampleTogether AI MoA · Self-MoA
When to use
Quality-critical tasks where you can trade cost/latency for a better answer, and want to avoid a single frontier dependency.
Basis
Verified against
https://arxiv.org/abs/2406.04692 The corpus does not rewrite vendor documentation. Where copy is quoted verbatim it is marked as such and attributed to its source.