Anthropic Engineer's Rule of Thumb: Each Sub-Agent Cuts Latency 50%, Doubles Tokens

Anthropic's pvncher shared a rule of thumb: each parallel sub-agent adds 100% token cost for 50% latency reduction. Community consensus: single agents are cheaper unless specific scenarios justify sub-agents.

2026-10-09 ~ 2026-10-09 · 3 related posts

• When sub-agents are actually worth it: fan-out research and pre-planned parallel tasks — brandon_galang · 2026-10-09
• Rule of thumb for sub-agents: ~50% latency gain per agent at 100% extra token cost — pvncher · 2026-10-09
• Rule of Thumb: Each Sub-agent Costs 2x Tokens for ~50% Latency Gain — brandon_galang · 2026-10-09