---
title: "Comment by Julian Neagu on \"Benchmark the Model–Harness Pair, Not the Model\" | Hashnode"
description: "The 40x token gap is the part that really stood out. A coding agent is only as good as the loop around the model."
url: "https://hashnode.com/posts/3-benchmark-the-model-harness-pair/6a8bf9b917de0e315d9dcde4/comment/6a8e85dbce5757b692289eba"
---

[Discussion](https://hashnode.com/posts/3-benchmark-the-model-harness-pair/6a8bf9b917de0e315d9dcde4)

Comment [Benchmark the Model–Harness Pair, Not the Model](https://hashnode.com/posts/3-benchmark-the-model-harness-pair/6a8bf9b917de0e315d9dcde4)

[Julian Neagu](https://hashnode.com/@julianneagulab)

[Aug 26](https://hashnode.com/posts/3-benchmark-the-model-harness-pair/6a8bf9b917de0e315d9dcde4/comment/6a8e85dbce5757b692289eba)

The 40x token gap is the part that really stood out. A coding agent is only as good as the loop around the model.

[Muhammad Azlaan Zubair](https://hashnode.com/@mdazlaanzubair)

[Aug 26](https://hashnode.com/posts/3-benchmark-the-model-harness-pair/6a8bf9b917de0e315d9dcde4/comment/6a8e85dbce5757b692289eba#reply-6a8e95bbc69942c3583976ee)

100% agree, [Julian](https://hashnode.com/@julianneagulab "Julian"). The model provides the reasoning, but the loop/harness defines whether it actually succeeds or breaks in production.

I recently wrote about this while analyzing leaked system prompts. Specifically looking at how production agents treat tool use as a hard operating contract (orchestration logic, failure handling, parallel execution) rather than simple prompt decoration:

[Are these really accidental leaks?](https://blog.mdazlaanzubair.com/are-these-really-accidental-leaks)

Without that explicit loop logic around the model, even the best model degrades quickly once edge cases hit.