# Technical Research - Conversational Dialog

**URL:** <https://forum.cursor.com/t/technical-research-conversational-dialog/170662>\
**Category:** Discussions\
**Created:** [September 5, 2026, 2:56am UTC](https://forum.cursor.com/t/technical-research-conversational-dialog/170662 "2026-09-05T02:56:50Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![ajaxStardust](https://sea3.discourse-cdn.com/cursor1/user_avatar/forum.cursor.com/ajaxstardust/32/103146_2.png) [@ajaxStardust](https://forum.cursor.com/u/ajaxStardust)\
**Post date:** [September 5, 2026, 2:56am UTC](https://forum.cursor.com/t/technical-research-conversational-dialog/170662/1 "2026-09-05T02:56:51Z")

</div>

Using Ollama cloud models (Kimi k2.7, GLM 5.2, Minimax) for discreet model per dialog, per turn, I made an Observation about Minimax regarding it’s tendency to use less cliche / stock patterns in conversational dialog.

Notes here: [Minimax is less cliche than other LLM's in a discourse — Wha](https://whatsonyourbrain.com/blog/minimax-less-cliche-other-llms-discourse)

Cheers!