A trace study models how findings survive through multi-agent delegation trees. Across 600 production research traces, each extra layer loses information, and one in sixteen annotated handoffs drifts off target. Two tiers beat a flat agent on cost only above roughly 403 findings under the measured parameters. A separate 743,819-tool-call analysis finds that production agents delegate early rather than in response to a filling context.
Study finds deep agent hierarchies preserve context but lose findings
2026-09-16Frontier watchTRIAL
A trace study models how findings survive through multi-agent delegation trees. Across 600 production research traces, each extra layer loses information, and one in sixteen annotated handoffs drifts off target. Two tiers beat a flat agent on cost only above roughly 403 findings under the measured parameters. A separate 743,819-tool-call analysis finds that production agents delegate early rather than in response to a filling context.
Imported from the existing research archive. Migration preserves the recorded judgment and is not a fresh verification.
What this means for you
For teams designing multi-agent workflows, hierarchy depth should be a measured budget decision rather than a default architecture. Keep fan-out wide, summaries explicit and depth small unless the task is large enough for context cost to dominate. The underlying production corpus is not released, so the numerical threshold needs local validation.
FRONTIER WATCH
Public artifacts and validated results are different.
The recorded fact is the public work and its artifacts. Author-reported outcomes do not establish independent reproduction. Next checks include usable code, task conditions and third-party validation.
Evidence & related material
Technical details & scores
| Research Traces | 600 |
|---|---|
| Annotated Handoffs | 1012 |
| Production Tool Calls | 743819 |
| Delegating Sessions | 7.8% of 7,244 sessions |
| Cost Crossover | two tiers overtake flat at about 403 findings |
| Limitation | single-author analysis; production corpus and code package are not linked publicly |
Technical impact 4
Engineering value 5
Adoption signal 2
Maturity 2
Verification cost 4
Risk 2