Models Can Model, But Can’t Bind: Structured Grounding in Text-to-Optimization

Text-to-optimization requires two separable capabilities: modeling—choosing the right optimization structure—and binding—grounding every coefficient, index, and parameter in the concrete problem data. We study this via OR-LLM-Bench, a scalable benchmark of solver-verified optimization problems spanning 12 categories, from textbook linear programs to stochastic and multi-objective formulations with up to thousands of variables. Across 10+ models, we find that accuracy collapses as instance data grows, even when the formulation itself is simple. We call this the effective binding limit. We address this via a simple inference-time approach, BIND, which externalizes numeric data to structured files so the model binds data programmatically rather than transcribing from the prompt. BIND improves GPT-5-Nano from 59.1\% to 82.4\% accuracy, matching pass@5 (82.0\%) at lower token cost than pass@1, and GPT-5 from 86.2\% to 95.8\%. Furthermore, we validate our hypothesis by finetuning a model exclusively on binding and show that it outperforms end-to-end SFT and RL across three structurally distinct optimization categories, with a 1.5B binding specialist alone matching a 7B end-to-end baseline.