A real-source data synthesis framework trains a 4B agent on 4k task instances with supervised fine-tuning and rubric-based reinforcement learning, improving realistic productivity work and file-grounded reasoning over larger models.
Large language models have made rapid progress in agentic reasoning and tool use, raising expectations for agents that can take over realistic productivity work from humans. However, current agents still struggle to finish such tasks, where success requires following long-horizon professional workflows, understanding heterogeneous file contexts, and producing usable deliverables under fine-grained requirements. A key bottleneck is the lack of scalable post-training data that reflects real work intent, authentic file environments, and deliverable-level evaluation. To close this gap, we introduce Agent-R^2, a real-source data synthesis framework that scales training data for realistic…
large language modelsagentic reasoningtool usepost-training datareal-source data synthesissupervised fine-tuningrubric-based reinforcement learningfile-grounded reasoningshortcut behaviorsustained exploration
- Upvotes
- 4
- Authors
- 8
- Published
- Nov 5, 2026
Summary and abstract excerpt via Hugging Face Papers ↗ · arXiv.