Qwen3.5-35B-A3B
#5
by E7Reine - opened
Good afternoon! I like your quant, they show themselves very well with a small weight in LM Studio.
Are you planning to create QWEN 3.5 35B A3B in the near future?
Hello!
Thank you for the kind words. There have been many interesting model releases lately, and Qwen3.5 is definitely among the best. It’s on our radar for future releases. :)
A "cpu" quant from byteshape will SAVE the using of this model on 16GB consumer cards. 3 bpw... chef's kiss
+1 for E7Reine's suggestion:
Qwen3.5-35B-A3B becomes very brittle at conventional quantizations below Q5_K_M. This would really be an intruiguing test case for the ByteShape approach!