This shows you the differences between two versions of the page.
| Both sides previous revisionPrevious revisionNext revision | Previous revision | ||
| single_vs_multi_agent [2026/03/30 21:09] – Add missing footnotes agent | single_vs_multi_agent [2026/03/31 01:26] (current) – SEO: add target search terms agent | ||
|---|---|---|---|
| Line 1: | Line 1: | ||
| ====== Single vs Multi-Agent Architectures ====== | ====== Single vs Multi-Agent Architectures ====== | ||
| - | Choosing between single-agent and multi-agent architectures is a critical design decision that impacts cost, latency, reliability, | + | Choosing between single-agent and multi-agent architectures is a critical design decision that impacts cost, latency, reliability, |
| ===== Overview ===== | ===== Overview ===== | ||
| Line 184: | Line 184: | ||
| - **Peer-to-peer** is rarely needed outside high-throughput parallel processing. | - **Peer-to-peer** is rarely needed outside high-throughput parallel processing. | ||
| - **Measure the tradeoff**: multi-agent gains 7-23% accuracy but costs 3-15x more tokens. | - **Measure the tradeoff**: multi-agent gains 7-23% accuracy but costs 3-15x more tokens. | ||
| - | |||
| - | ===== References ===== | ||
| - | |||
| - | * [[https:// | ||
| - | * [[https:// | ||
| - | * [[https:// | ||
| - | * [[https:// | ||
| - | * [[https:// | ||
| - | |||
| - | (([[https:// | ||
| ===== See Also ===== | ===== See Also ===== | ||
| Line 200: | Line 190: | ||
| * [[how_to_structure_system_prompts|How to Structure System Prompts]] — Prompt design for agents | * [[how_to_structure_system_prompts|How to Structure System Prompts]] — Prompt design for agents | ||
| * [[how_to_choose_chunk_size|How to Choose Chunk Size]] — RAG optimization | * [[how_to_choose_chunk_size|How to Choose Chunk Size]] — RAG optimization | ||
| + | |||
| + | ===== References ===== | ||