推文

@Michaelzsguo · 2026-10-11 04:09

If you're looking for a serious Jev-style decision model you can run locally, check out Quyet-1.0-Large. It tops the Jev Bench leaderboard. Quyet-1.0-Large is a 31.3B-parameter decision model built by fine-tuning Google's Gemma 4 31B Instruct with LoRA. I've since switched my Muse agent to Quyet (8-bit), running locally.

曝光 298 · 评论 3 · 点赞 3 · 书签 4 · 曝光/时 23.42150220671457