I'm convinced we will get Qwen4-flash as an open weight model.
I don't even think even
@Alibaba_Qwen imagined how much their preview model Qwen3.8-flash-next will be adopted by Local AI consumers, it's running basically everywhere.
It's the goto model on my 3090 now.
Released as a preview for Qwen4 architecture, giving us time to adopt and develop tooling around it. It took some time but now we have support everywhere.
- GSQ-RSO quant got the size down without much quality loss.
- Strata arrived and we all know what happened next.
Most people who relied on 35b are now running it. They finally got the upgrade they were waiting for.
The community has put so much work into making this model accessible. Imagine what we could do with Qwen4-Flash.
So
@Alibaba_Qwen, take some time, but def do release Qwen4-flash open weight.
We're ready for it.