# Gorgon Halo + DeepSeek V4.1-Flash-1017 + Maya

**URL:** <http://gitLair.com/t/gorgon-halo-deepseek-v4-1-flash-1017-maya/104>\
**Category:** AI & Automation\
**Created:** [November 1, 2026, 4:35am UTC](http://gitLair.com/t/gorgon-halo-deepseek-v4-1-flash-1017-maya/104 "2026-11-01T04:35:26Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![kunlun](http://gitLair.com/user_avatar/gitLair.com/kunlun/32/45_2.png) [@kunlun](http://gitLair.com/u/kunlun)\
**Post date:** [November 1, 2026, 4:35am UTC](http://gitLair.com/t/gorgon-halo-deepseek-v4-1-flash-1017-maya/104/1 "2026-11-01T04:35:26Z")

</div>

Gorgon Halo + DeepSeek V4.1-Flash-1017 + Maya

Are we there yet?

763B Parameters model ✓  
Sustained decode speeds ✓  
Very high prefill ✓  
Context dept doesn’t affect speeds ✓  
Model recommended parameters, NEVER greedy decode ✓

 ![HV_1MyOWwAAaaQ4](http://gitLair.com/uploads/default/original/1X/62ba3e28f24db59a5d5d8c68136aeae8c80232b4.jpeg)