Hacker News new | ask | show | jobs
by v3ss0n 29 days ago
3.5 122B is much better. 27 B is bad at Long context and Svelte