Prefer Kimi-K3 Over Grok 4.6: 'Big Model Smell' Makes a Difference

scaling01 · x · 2026-08-13

Continuing the discussion on his personal LLM tier list, the author states that despite Grok 4.6 being potentially underrated, he would still prefer using Kimi-K3.

He argues that the 'big model smell' (overall vibe and nuance) makes a significant difference in practical use. He also questions whether Grok 4.6 should objectively be considered better than GPT-5.4 and Opus 4.7.

Original post →

More from Models

Models channel →