We need GLM-5 and Qwen 3.5(397B and 122B) ASAP!

So many people are interested in these newest powerful models. Would be so bad to migrate to other fast-inference providers that will have it first.