Text2Model Release

[Serdar Kadioglu]

Hello everyone,

It is my pleasure to announce the release of 

Text2Model available via `pip install text2model’

Text2Model is a suite of modeling copilots, dataset, editor, and fined-tuned models to convert problem descriptions into MiniZinc models. 

• Strategies: It provides several strategies including zero-shot, chain-of-thought, knowledge graphs, (global) agentic workflows, grammar encodings, etc. 
  `text2model --strategies’

• Usage Modes: It can work with your free-from text `text2model --problem’ or with text2zinc instances `text2model --problem-ids’

• Editor: The interactive editor `text2model --editor’ allows data inspection and curation

• SLM: We also provide MiniZinc fined-tuned versions of Qwen, Llama, Gemini, and GPT-OSS to support local development environments.

We look forward to your feedback and contributions. 

Best,

Serdar

Ps: A tutorial on this line of research is upcoming soon at IJCAI’26

https://sites.google.com/view/ijcai26-llm-con