Link model to NexForge paper and project page

This PR updates the model card to link it to the official paper [NexForge: Scaling Executable Agent Tasks via Requirement-First Synthesis](https://arxiv.org/abs/2607.14186) and adds links to the project page, as well as a citation section at the end of the README. This helps users understand the context and methodology behind the Nex-N2 model family and cite your work correctly.

README.md

@@ -1,7 +1,7 @@
 ---
+library_name: transformers
 license: apache-2.0
 pipeline_tag: text-generation
-library_name: transformers
 ---
 
 <div align="center">
@@ -13,6 +13,8 @@ library_name: transformers
 <div align="center">
 🤗 <a href="https://hf.co/collections/nex-agi/nex-n2"><b>Model</b></a>&nbsp&nbsp | &nbsp&nbsp
 🔀 <a href="https://openrouter.ai/nex-agi/Nex-N2-Pro:free"><b>OpenRouter (Enjoy two weeks free starting June 9!)</b></a>&nbsp&nbsp | &nbsp&nbsp
+📄 <a href="https://arxiv.org/abs/2607.14186"><b>Paper</b></a>&nbsp&nbsp | &nbsp&nbsp
+🌐 <a href="https://nex.sii.edu.cn/"><b>Project Page</b></a>&nbsp&nbsp | &nbsp&nbsp
 💻 <a href="https://github.com/nex-agi/Nex-N2"><b>Github</b></a>&nbsp&nbsp | &nbsp&nbsp
 🧭 <a href="https://www.modelscope.cn/collections/nex-agi/Nex-N2"><b>ModelScope</b></a>&nbsp&nbsp | &nbsp&nbsp
 🚀 <a href="https://nex-agi.com"><b>Nex-AGI</b></a>
@@ -22,6 +24,8 @@ library_name: transformers
 
 **An agentic model with Agentic Thinking.**
 
+Nex-N2 is trained using data synthesized by **NexForge**, a requirement-first framework that compiles free-form capability requirements into executable agent training data. For more details on the synthesis process and evaluation results, please refer to the paper: [NexForge: Scaling Executable Agent Tasks via Requirement-First Synthesis](https://arxiv.org/abs/2607.14186).
+
 Today, we are officially releasing and open-sourcing our next-generation model, **Nex-N2** — an agent model built for real-world productivity scenarios. With first-tier coding and agentic capabilities, Nex-N2 keeps driving complex, long-horizon tasks forward in real environments to deliver stable, end-to-end results.
 
 Over the past year, a paradigm shift led by Vibe Coding and Harness Engineering has been redefining the limits of LLM agents. From dialogue, to reasoning, to agents that execute long-horizon tasks with environmental feedback, the tasks models must handle keep growing harder, the contexts longer, and the environments more realistic. The core of next-generation model competition is no longer *whether a model can think*, but whether it can reliably and efficiently turn thinking into actions that are executable, verifiable, and iterable.
@@ -189,3 +193,19 @@ Nex-series models emit explicit reasoning traces. Add the `--reasoning-parser qw
 ```bash
 python -m sglang.launch_server --model-path /path/to/your/model --tool-call-parser qwen3_coder --reasoning-parser qwen3
 ```
+
+## Citation
+
+If you find our work or the NexForge framework useful, please cite our paper:
+
+```bibtex
+@misc{zhao2026nexforge,
+  title={NexForge: Scaling Executable Agent Tasks via Requirement-First Synthesis}, 
+  author={Jiarong Zhao and Zhikai Lei and Zhiheng Xi and Rui Zheng and Hang Yan and Jie Zhou and Qin Chen and Liang He},
+  year={2026},
+  eprint={2607.14186},
+  archivePrefix={arXiv},
+  primaryClass={cs.CL},
+  url={https://arxiv.org/abs/2607.14186}
+}
+```