DeepSeek V4 Flash: 284B model, up to 32 tok/s on AMD Ryzen AI MAX+ 395

Anyone try this out yet?

2 Likes

I was able to get Deepseek V4 flash running on my framework desktop - just as an experiment. It is limiting on the context size, but it does actually work which is pretty cool.

I will stick to qwen3.6 for daily use however.

1 Like