@MiaAI_lab: Flint 40B Is Fast, but It Can’t Replace Qwen3.6-35B

Fast is not the same as better, says @MiaAI_lab.

@MiaAI_lab tested Flint 40B: “It’s very fast. Unfortunately, it’s not, in any way, shape or form, a replacement for Qwen3.6-35B.” In a later post, @MiaAI_lab asks when a truly worthy local model will beat “the legendary Qwen3.6-35B”: this year, 2027 or never.

Separately, GLM 5.3 Flash now runs on TensorFold ZIG on all configurations: TP2, TP3 and TP4, with or without a switch. It is on a branch separate from main, which stays on Python until everything is stable. @MiaAI_lab thanks @hecisaza for the work.