3 ms·
on the Will It Mythos benchmark, small models are punching way above their weight(s) gemma4-26B (#7) qwen-3.6-27B (#9) https://news.ycombinator.com/item?id=4
by verdverm 4mo ago
on the Will It Mythos benchmark, small models are punching way above their weight(s)
gemma4-26B (#7)
qwen-3.6-27B (#9)
https://news.ycombinator.com/item?id=48640196 https://news.ycombinator.com/item?id=48640196
- Gigachad 4mo agoI've tried running qwen 3.6 locally and it felt like LLMs a year ago where you can get them to do some stuff but the tasks have to be very small and you have to course correct them a lot to the point it's hard to say it's any faster than doing it all yourself. Certainly the gap is closing but I feel it still makes more sense to pay pennies to run the full sized open models hosted on much better hardware.
- verdverm 4mo agoI had qwen36moe revamp my PhD thesis with a rewrite using JAX. Gave it access to my old code, helpednitnwhen it got stuck or didn't quite understand a few times. Overall I was very impressed with its open box reimplementation. I remain of the mind they are widely underrated.
- trollbridge 4mo agoWhat edition of Qwen 3.6? 35b-q6 (with MOE) has felt good enough for general purpose agentic coding. You obviously need a 32GB card to run that (or a 64GB Mac), and realistically anything lesser than an RTX 5090 is going to be too slow for practical use.