Agents
Modeling Co-Pilots for Text-to-Model Translation
arXiv:2604.12955v1 Announce Type: new Abstract: There is growing interest in leveraging large language models (LLMs) for text-to-model translation and optimization tasks. This paper aims to advance th
arXiv:2604.12955v1 Announce Type: new Abstract: There is growing interest in leveraging large language models (LLMs) for text-to-model translation and optimization tasks. This paper aims to advance this line of research by introducing extsc{Text2Model} and extsc{Text2Zinc}. extsc{Text2Model} is a suite of co-pilots based on several LLM strategies with varying complexity, along with an online leaderboard. extsc{Text2Zinc} is a cross-domain dataset for capturing optimization and satisfaction problems specified in natural language, along with an interactive editor with built-in AI assistant. While there is an emerging literature on using LLMs for translating combinatorial problems into formal models, our work is the first attempt to integrate extit{both} satisfaction and optimization problems within a extit{unified architecture} and extit{dataset}. Moreover, our approach is extit{solver-agnostic} unlike existing work that focuses on translation to a solver-specific model. To achieve this, we leverage extsc{MiniZinc}'s solver-and-paradigm-agnostic modeling capabilities to formulate combinatorial problems. We conduct comprehensive experiments to compare execution and solution accuracy across several single- and multi-call strategies, including; zero-shot prompting, chain-of-thought reasoning, intermediate representations via knowledge-graphs, grammar-based syntax encoding, and agentic approaches that decompose the model into sequential sub-tasks. Our co-pilot strategies are competitive, and in parts improve, recent research in this domain. Our findings indicate that while LLMs are promising they are not yet a push-button technology for combinatorial modeling. We contribute extsc{Text2Model} co-pilots and leaderboard, and extsc{Text2Zinc} and interactive editor to open-source to support closing this performance gap.
Related
- Model Space Reasoning as Search in Feedback Space for Planning Domain Generation
- ALL-FEM: Agentic Large Language models Fine-tuned for Finite Element Methods
- Automating Structural Analysis Across Multiple Software Platforms Using Large Language Models
- Towards Autonomous Mechanistic Reasoning in Virtual Cells
Source: arXiv cs.AI | 2026-04-15