Abstract

Coding-agent harnesses are usually compared by model quality or task speed. This study asks a more physical question: what does each installed harness cost the local system while it starts and while it works? OpenCode, Codex CLI, Cursor Agent CLI, Pi, Grok Build, and fx were measured on a Linux workstation using fresh process trees, fixed coding tasks, and explicit solo and delegated arms.

1.47 MB

fx cold-start RSS

Lowest median startup footprint

240.6 MB

Pi light-task RSS

Lowest matched Luna-max solo task tree

12.92 s

Codex heavy CPU

Lowest matched Luna-max solo CPU time

220.4 MB

OpenRouter Grok RSS

Solo heavy-task supplement median

Quick read

The result at a glance

Bars show medians. Lower is lighter or less CPU-intensive.

Cold-start peak RSS

MB

Lower is lighter before model inference begins.

Heavy task: Luna-max solo RSS

MB

The closest apples-to-apples active-task comparison.

Heavy task: Luna-max solo CPU

CPU s

Total local user plus system CPU seconds; lower is less CPU-intensive.

Short version

There is no single winner. fx is the smallest cold start, Pi is the lightest matched Luna-max task tree, Codex uses the least heavy-task CPU, Cursor finishes the matched heavy task fastest, and Grok Build with OpenRouter Grok 4.6 is the strongest complete deployment observed in the supplement.

01

The question

The study separates cold startup from active deployment. A fresh CLI process answers how much local work a harness performs before inference. A task process tree answers what the harness, its tools, subprocesses, and model-driven workflow cost together.

That distinction matters. A small native binary can still become expensive once context, tools, and delegated workers are active. Conversely, a heavier CLI can complete a task quickly if its workflow is efficient.

02

What the measurements show

The matched Luna-max subset is the closest comparison between harnesses. Pi had the lowest peak process-tree RSS on both task sizes. Codex used the least CPU on the heavy task. Cursor completed that heavy task fastest. OpenCode produced the largest repeated task tree in this setup.

A separate OpenRouter supplement kept the Grok Build harness but routed it to x-ai/grok-4.6. All twelve supplement attempts completed, separating the earlier native Grok quota failure from the harness itself.

03

How it was measured

Twenty fresh version/help trials per harness measured startup. Three trials per arm/task cell measured deployment. Every task started from a fresh fixture copy and ran sequentially on one Linux x86-64 machine with four logical CPUs and approximately 15.5 GiB of RAM.

A 20 ms /proc sampler followed the complete descendant process tree. It recorded summed RSS, CPU seconds, process and thread peaks, context switches, and I/O. GNU time supplied a fallback for native processes that exited before a sampler tick.

  • Light task: an eight-check retry-with-backoff maintenance issue.
  • Heavy task: three independent implementation, security, and comprehension packages with 24 hidden checks.
  • Solo arms disabled delegation where possible; delegated arms enabled each harness-specific mechanism.
04

What it does not prove

This is a deployment-envelope comparison, not a universal causal ranking of harness implementations. Remote models can choose different turn counts, tool calls, and child processes, and those choices change local resource use even when the model names look similar.

The task sample is small and synthetic, provider-side compute is invisible, and Grok native task repetition was quota-censored. The most defensible reading is practical: these are the resource envelopes observed for these versions, policies, models, and tasks.

Cold startup
HarnessPeak RSSCPU
fx1.47 MB<0.01 s
Grok Build27.0 MB0.02 s
Codex CLI50.5 MB0.09 s
Pi157.9 MB2.51 s
OpenCode187.5 MB1.85 s
Cursor Agent206.4 MB1.76 s
Matched Luna-max solo tasks
HarnessLight RSSHeavy RSSHeavy wall
Pi240.6 MB535.8 MB1047.5 s
Cursor363.4 MB565.0 MB458.3 s
Codex CLI441.1 MB608.9 MB851.7 s
OpenCode742.8 MB899.6 MB836.5 s
Grok Build with OpenRouter Grok 4.6
ArmTaskPeak RSSScore
SoloLight213.9 MB8/8
DelegatedLight163.6 MB8/8
SoloHeavy220.4 MB23/24
DelegatedHeavy307.7 MB23/24

Research note

This is a Katteke working paper, not an official arXiv submission or peer-reviewed publication. The source report, data, and replication notes are linked below so the claims can be inspected in context.