Rufus-Air: An Open LLM Post-Training Recipe
Rufus-Air releases a reproducible eight-stage post-training recipe on GLM-4.5-Air-Base that beats the official post-trained release.
Rufus-Air documents an open post-training pipeline on GLM-4.5-Air-Base (106B-A12B) consisting of eight serial stages: SFT, Reasoning RL, Coding RL, Instruction-Following RL, General Agent, Coding Agent, Search Agent, and RLHF. It uses only open-source components and public data with no new human annotation or in-house distillation teacher. Key findings include that diverse high-quality SFT establishes a strong capability floor, difficulty filtering keeps RL prompts productive, and reward reliability guides stage ordering. The result improves over the official GLM-4.5-Air post-trained release and is competitive with similarly sized open models.
- Eight-stage recipe: SFT, three RL capability stages, three agent stages, then RLHF
- Built entirely on public data and open-source infrastructure, no new human annotation
- Outperforms official GLM-4.5-Air post-trained release; competitive at its size class
Full article157 words · extracted from huggingface.co · click to collapse
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 Agent, Search Agent, and RLHF. We document the data, reward design, infrastructure, stage order, and stagewise results needed to reproduce the recipe. Stages progress from basic to advanced capabilities and from hard, verifiable rewards to softer judge-based signals. Training builds on open-source components and public data, much of it used as released, without new human annotation or an in-house distillation teacher. Our main findings are that (i) diverse, high-quality SFT establishes a strong capability floor; (ii) difficulty filtering keeps RL prompts within a productive learning range; (iii) reward reliability provides a practical principle for ordering stages; and (iv) infrastructure and engineering choices are part of the recipe, not just an implementation detail. Rufus-Air improves over the official GLM-4.5-Air post-trained release and is competitive with similarly sized open models.
Text extracted automatically; images, tables and formatting may be missing. Original: https://huggingface.co/papers/2609.29421