5 ms·
https://modelscope.cn/models/Qwen/Qwen3.8-27B/summary https://modelscope.cn/models/Qwen/Qwen3.8-27B/summary Countdown at 28h now. Sadly they seem to not be rel
by underlines 1mo ago
https://modelscope.cn/models/Qwen/Qwen3.8-27B/summary https://modelscope.cn/models/Qwen/Qwen3.8-27B/summary
Countdown at 28h now.
Sadly they seem to not be releasing a sparse 35b A3b or anything inbetween "too large to host for mortals" and "fits into a consumer rtx". Probably not to eat away their profits on their API serving. 120b - 300b is a dead space right now, very few good releases in that size range. (I know there are, but the big labs aren't releasing stuff here)