AI article
We ran 2 vs 4 agents six times. Four agents cost 2.1 and did not improve success
Community description: A preregistered deterministic task, six valid live runs, hard call and cost caps, and a negative result. More agents were operationally viable but did not improve task success, at 2.115× the cost.
Dev.to | Sep 22, 2026 | Janz
Automated excerpt
The 4-agent group cost 2. 115× more per run and per complete success. The sample is three repeats per group on one synthetic task. The result is intentionally boring: a negative result with caps, hashes, and every failed run preserved.
Selected automatically from source text; not independently written or fact-checked. Read the original for full context.