Paper Detail

Agent-R^2: Scaling Agents for Realistic Productivity Work via Real-Source Data Synthesis

Teng Pan, Zixuan Wang, Yuchen Yan, Qifei Wu, Yao Chen, Yingqi Qu, Ruiqing Zhang, Yongliang Shen

huggingface Score 22.4

Published 2026-11-05 · First seen 2026-11-13

General AI

Abstract

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 productivity work. Instead of relying on manually specified workflows or LLM-generated files, Agent-R^2 grounds synthesis throughout the pipeline in real occupational seeds, corpus-derived workflow evidence, and authentic web-sourced files. From 10k web-corpus candidates, Agent-R^2 distills 4k task instances that pair real multi-format attachments with fine-grained, deliverable-level rubrics. On this data, we train Agent-R^2-4B via supervised fine-tuning and rubric-based reinforcement learning, improving both end-to-end productivity performance and file-grounded reasoning. Trained on only 4k instances, Agent-R^2-4B surpasses Qwen3-235B-A22B-Thinking-2507 on both GDPval (51.0% vs. 48.0%) and OfficeQA (14.8% vs. 13.8%). Further analysis shows that Agent-R^2 mitigates shortcut behavior, where agents produce superficially plausible deliverables before sufficiently grounding, constructing, or verifying them, while promoting sustained exploration.

Workflow Status

Review status
pending
Role
unreviewed
Read priority
now
Vote
Not set.
Saved
no
Collections
Not filed yet.
Next action
Not filled yet.

Reading Brief

No structured notes yet. Add `summary_sections`, `why_relevant`, `claim_impact`, or `next_action` in `papers.jsonl` to enrich this view.

Why It Surfaced

No ranking explanation is available yet.

Tags

No tags.

BibTeX

@misc{pan2026agent,
  title = {Agent-R\textasciicircum{}2: Scaling Agents for Realistic Productivity Work via Real-Source Data Synthesis},
  author = {Teng Pan and Zixuan Wang and Yuchen Yan and Qifei Wu and Yao Chen and Yingqi Qu and Ruiqing Zhang and Yongliang Shen},
  year = {2026},
  abstract = {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 i},
  url = {https://huggingface.co/papers/2611.08219},
  keywords = {large language models, agentic reasoning, tool use, post-training data, real-source data synthesis, supervised fine-tuning, rubric-based reinforcement learning, file-grounded reasoning, shortcut behavior, sustained exploration, GDPval, OfficeQA, Qwen3-235B-A22B-Thinking-2507, code available, huggingface daily},
  eprint = {2611.08219},
  archiveprefix = {arXiv},
}

Metadata

{}