[–] nshotton 1mo ago ↗ you can see the countdown here: https://modelscope.cn/models/Qwen/Qwen3.8-2.4T-A95B?spm=a2ty...
[–] wing-_-nuts 1mo ago ↗ I love these small models that are able to fit onto local hardware. Even they are capable enough to be helpful and while I think AI is still a bubble, these open models give me faith that the underlying tech isn't going away.
[–] t3hTao 1mo ago ↗ As an owner of a company spending $4000/mo on token spend. Yawn...I tried every model under 1T and they are not useful for AI Agents.If I want to use a regular chat model about amoral actions, the 7-9B abliterated models were fine.I'm sure these 27B models can get you the weather after 5 minutes for the people who didn't buy Nvidia chips.
[–] pivi 1mo ago ↗ maybe there's some kind of a delay that they have to remove the countdown. I saw it too when it had the countdown and now it's 404
10 comments
[ 0.25 ms ] story [ 12.0 ms ] threadI tried every model under 1T and they are not useful for AI Agents.
If I want to use a regular chat model about amoral actions, the 7-9B abliterated models were fine.
I'm sure these 27B models can get you the weather after 5 minutes for the people who didn't buy Nvidia chips.