arXiv.org
arXiv.org is Rufus-Air is an open and reproducible post-training recipe on GLM-4.5-Air-Base (106B-A12B), organized as a serial pipeline of eight stages: SFT, Reasoning RL, Coding RL, Instruction-Following RL, General Agent, Coding…. It is ranked #397 on the AI Radar, in AI research, first seen 6 h ago and shared in 1 post (3.3k views).
[2609.29421] Rufus-Air: An Open LLM Post-Training Recipe Skip to main content Search Search arXiv Press Enter to search · Advanced search --> Computer Science > Computation and Language arXiv:2609.29421 (cs) [Submitted on 24 Sep 2026 ( v1 ), last revised 25 Sep 2026 (this version, v2)] Title: Rufus-Air: An Open LLM Post-Training Recipe Authors: Chia-Yuan Chang , Renyuan Cheng , Rui Feng , Xiaotian Han , Yuan He , Hongye Jin , Linwei Li , Shiyang Li , Fenglin Liu , Xin Liu , Priyanka Nigam , Haoyang Wen , Zhenghao Xu , Zhuocheng Xu , Bing Yin , Qingyu Yin , Chao Zhang , Rongzhi Zhang , Zhihan Zhang , Zixuan Zhang , Zixuan Zhang , Tuo Zhao…
What people said about arXiv.org on X
🔥我靠！终于有大公司把AI模型后训练的“魔法”拆成可复现配方了！ Amazon团队直接公开从GLM-4.5-Air-Base出发的完整8阶段串行后训练pipeline，数据来源、奖励设计、筛选方法、训练配置、阶段结果全部写清楚，且大量用公开数据、无新增人工标注、无内部蒸馏教师！ 他们发现： 1️⃣串行8阶段流水线（每一步继承上一步参数）： SFT → Reasoning RL → Coding RL → Instruction-Following RL → General Agent → Coding Agent → Search Agent → RLHF 从基础能力到高级Agent，从硬可验证奖励逐步过渡到软judge奖励。…
— @NFT_Chen, 6 h ago · 53 likes · see the post
Alternatives to arXiv.org
- alphaXiv — Now on @askalphaxiv
- flashprove — Further progress in provable inference will likely be an important part of improving trust in AI deployment. To that…
- moe-blogpost — A hands-on guide to training Mixture-of-Experts models
- liberai.com — Technical blog
- Science's Last Exam — Can AI do research, end to end? Science's Last Exam defines and measures whether an AI agent can carry a research goal…
- science-last-exam — Science's Last Exam. Contribute to Gen-Verse/Science-Last-Exam development by creating an account on GitHub.
arXiv.org in numbers
- Rank on the AI Radar: #397 of 10395
- Shared in 1 post by 1 account: @NFT_Chen
- 3.3k views on those posts
- First seen 6 h ago, last shared 6 h ago
- Pricing seen by AGK Intelligence: open source
- Market: AI research
FAQ
What is arXiv.org?
Rufus-Air is an open and reproducible post-training recipe on GLM-4.5-Air-Base (106B-A12B), organized as a serial pipeline of eight stages: SFT, Reasoning RL, Coding RL, Instruction-Following RL, General Agent, Coding… It was first shared on X 6 h ago and is ranked #397 on the AI Radar.
Is arXiv.org free?
It is open source.
Who shared arXiv.org?
1 account on X, including @NFT_Chen, in 1 post totalling 3.3k views.
Every post is read and classified by AGK Intelligence: what it is, which market it belongs to, and whether the link is a real tool. 65.6k posts from 5k X accounts over the last 14 days, 10.4k tools, 19 markets. New posts are collected several times an hour and the full ranking is recomputed every hour — last update 2026-11-06 14:15 UTC. Full method.