Blueprint First, Model Second: A Framework for Deterministic LLM Workflow
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866912519496400896 |
|---|---|
| author | Qiu, Libin Ye, Yuhang Gao, Zhirong Zou, Xide Chen, Junfu Gui, Ziming Huang, Weizhi Xue, Xiaobo Qiu, Wenkai Zhao, Kun |
| author_facet | Qiu, Libin Ye, Yuhang Gao, Zhirong Zou, Xide Chen, Junfu Gui, Ziming Huang, Weizhi Xue, Xiaobo Qiu, Wenkai Zhao, Kun |
| contents | While powerful, the inherent non-determinism of large language model (LLM) agents limits their application in structured operational environments where procedural fidelity and predictable execution are strict requirements. This limitation stems from current architectures that conflate probabilistic, high-level planning with low-level action execution within a single generative process. To address this, we introduce the Source Code Agent framework, a new paradigm built on the "Blueprint First, Model Second" philosophy. Our framework decouples the workflow logic from the generative model. An expert-defined operational procedure is first codified into a source code-based Execution Blueprint, which is then executed by a deterministic engine. The LLM is strategically invoked as a specialized tool to handle bounded, complex sub-tasks within the workflow, but never to decide the workflow's path. We conduct a comprehensive evaluation on the challenging tau-bench benchmark, designed for complex user-tool-rule scenarios. Our results demonstrate that the Source Code Agent establishes a new state-of-the-art, outperforming the strongest baseline by 10.1 percentage points on the average Pass^1 score while dramatically improving execution efficiency. Our work enables the verifiable and reliable deployment of autonomous agents in applications governed by strict procedural logic. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2508_02721 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Blueprint First, Model Second: A Framework for Deterministic LLM Workflow Qiu, Libin Ye, Yuhang Gao, Zhirong Zou, Xide Chen, Junfu Gui, Ziming Huang, Weizhi Xue, Xiaobo Qiu, Wenkai Zhao, Kun Software Engineering Artificial Intelligence Programming Languages While powerful, the inherent non-determinism of large language model (LLM) agents limits their application in structured operational environments where procedural fidelity and predictable execution are strict requirements. This limitation stems from current architectures that conflate probabilistic, high-level planning with low-level action execution within a single generative process. To address this, we introduce the Source Code Agent framework, a new paradigm built on the "Blueprint First, Model Second" philosophy. Our framework decouples the workflow logic from the generative model. An expert-defined operational procedure is first codified into a source code-based Execution Blueprint, which is then executed by a deterministic engine. The LLM is strategically invoked as a specialized tool to handle bounded, complex sub-tasks within the workflow, but never to decide the workflow's path. We conduct a comprehensive evaluation on the challenging tau-bench benchmark, designed for complex user-tool-rule scenarios. Our results demonstrate that the Source Code Agent establishes a new state-of-the-art, outperforming the strongest baseline by 10.1 percentage points on the average Pass^1 score while dramatically improving execution efficiency. Our work enables the verifiable and reliable deployment of autonomous agents in applications governed by strict procedural logic. |
| title | Blueprint First, Model Second: A Framework for Deterministic LLM Workflow |
| topic | Software Engineering Artificial Intelligence Programming Languages |
| url | https://arxiv.org/abs/2508.02721 |