---
title: "Comment by Mihai_LeanZero on \"Does Size Matter? Bigger AI Isn’t Always Better\" | Hashnode"
description: "Worth pushing on the 405B-vs-Pythia example a bit further: that LLaMA was 4-bit quantized while the Pythia models weren't, and outlier features (the ones LLM.in"
url: "https://hashnode.com/posts/does-size-matter-bigger-ai-isn-t-always-better/6aaf0e4f4a3380a14d6989c7/comment/6aaf38148576fd0de007aa25"
---

[Discussion](https://hashnode.com/posts/does-size-matter-bigger-ai-isn-t-always-better/6aaf0e4f4a3380a14d6989c7)

Comment [Does Size Matter? Bigger AI Isn’t Always Better](https://hashnode.com/posts/does-size-matter-bigger-ai-isn-t-always-better/6aaf0e4f4a3380a14d6989c7)

[Mihai\_LeanZero](https://hashnode.com/@mihaileanzero)

[Sep 20](https://hashnode.com/posts/does-size-matter-bigger-ai-isn-t-always-better/6aaf0e4f4a3380a14d6989c7/comment/6aaf38148576fd0de007aa25)

Worth pushing on the 405B-vs-Pythia example a bit further: that LLaMA was 4-bit quantized while the Pythia models weren't, and outlier features (the ones LLM.int8() had to special-case) get more pronounced as models scale up. Naive quantization degrades a big model's effective capability more than it does a small one's, so there's a second confound sitting right on top of the ones already listed — a fair size comparison needs matched precision, not just matched task. Ran into this exact trap comparing an MXFP8 128GB local model against smaller full-precision ones; had to control for the quantization scheme before any size conclusion was trustworthy.

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

[Sep 20](https://hashnode.com/posts/does-size-matter-bigger-ai-isn-t-always-better/6aaf0e4f4a3380a14d6989c7/comment/6aaf38148576fd0de007aa25#reply-6ab0252d072ecca869081910)

Quantizing a 405B model to 4-bit precision while maintaining full precision for smaller models is an unfair comparison due to outlier features and model size.