/AI8h ago

Developer Switches From Qwen 35B Quant To Gemma 12B For Local Coding

--0--
Original posts
merve#674
Comments
merve#674
Original post
merve@mervenoyann#674inAI

just replaced my Qwen3.6 35B 8-bit quant with Gemma 12B bf16 for local coding & Hermes, will report my findings 馃檶馃徎

3:56 AM 路 Jun 4, 2026 路 53.2K Views
Sentiment
Sentiment building, check back later.
Cluster Engagement
-
Views
-
Comments
-
Reposts
-
Bookmarks
Expand data
Posts from X
Most Activity
Most ActivityTimeline
VIEWS1.5KBOOKMARKS11LIKES17RETWEETS1REPLIES1
merve@mervenoyann

if you are running Gemma-4 with Hermes/Claw, you need to modify the functions in the harness' message prep according to modality order

Gemma-4 takes in image/audio/text in certain order 猡碉笍 move text after image or ask your agent to do it across the library

merve@mervenoyann

just replaced my Qwen3.6 35B 8-bit quant with Gemma 12B bf16 for local coding & Hermes, will report my findings 馃檶馃徎

2hViews 1.5KLikes 17Bookmarks 11