AI article
I Pre-Registered a Four-Way LangGraph Shootout. Plan-and-Execute Cost 2.3x and Bought Nothing I Could Detect
Community description: Four LangGraph agent designs, one pre-registered question set: plan-and-execute used 2.3x the tokens of plain ReAct with no detectable accuracy gain (synthetic questions, one 9B model).
Dev.to | Oct 17, 2026 | Raihan
Automated excerpt
Four agent designs, the same 240 questions about Japanese statutes, the same local 9B model and the same token budget. "Correct" means the answer cites every gold article and no article that does not exist. Draft-verify was ahead of ReAct by 10 questions (16 only it got right, 6 only ReAct).
Selected automatically from source text; not independently written or fact-checked. Read the original for full context.