@ahmadmanga's thread

With how good Qwen3.5 small models are, I won't be surprised if by the end of the year we get an amazing MoE model that rivals the biggest models of 2025 with just 1B of active parameters.

Qwen3.5's 4B model is better AND faster than the similar sized Qwen3's 4B model... And the 0.8B model is able to do things that were previously unthinkable with that size! #ai #technology #qwen