2 ms·
It's just another Qwen 3.6 27B finetune (not even 3.8...), claiming to beat TB-sized models. This is probably the 57th model like that, just this week, claiming
by Karuma 24d ago
It's just another Qwen 3.6 27B finetune (not even 3.8...), claiming to beat TB-sized models. This is probably the 57th model like that, just this week, claiming the same thing... And they never deliver, to say the least.
- dgellow 24d agoDidn’t they deliver something here? Why does it matter what base model they used?
- eightysixfour 24d agoWhat they all deliver is benchmaxxed models that are awful in actual use.
- xena 24d agoBut it gives the lab good publicity, as seen in the article.
- dgellow 24d agoQwen is a really awesome model to use!
- vkaku 24d agoI'm giving out more unpopular opinions today for entertainment and thought :) Every model is benchmaxxing today. Many of them are awful until tuned for any particular task.
- alansaber 24d agoThe finetunes will continue until morale increases.
- BoredomIsFun 24d ago> (not even 3.8...) 3.6 is better for non-coding tasks, noticeably so.