We evaluate language models on their ability to explore interactive environments under a limited interaction budget. We introduce three parametric tasks with controllable exploration difficulty, spanning continuous and discrete environments. Across state-of-the-art models, we find systematic under-exploration and suboptimal solutions, with performance often significantly worse than simple explore--exploit heuristic baselines and scaling weakly as the budget increases. Finally, we study two lightweight interventions: splitting a fixed budget into parallel executions, which surprisingly improves performance despite a no-gain theoretical result for our tasks, and periodically summarizing the interaction history, which preserves key discoveries and further improves exploration.
@article{arxiv.2601.22345,
title = {Failing to Explore: Language Models on Interactive Tasks},
author = {Mahdi JafariRaviz and Keivan Rezaei and Arshia Soltani Moakhar and Zahra Sodagar and Yize Cheng and Soheil Feizi},
journal= {arXiv preprint arXiv:2601.22345},
year = {2026}
}