LATIDIA · Robótica

Los LLM de Frontier entran en el mundo de la robótica: una evaluación de modelos cruzados en RoboColiseum

arXiv: 2610.27865v1Tipo de anuncio: nuevo Resumen: Estudios recientes han explorado el potencial de los modelos multimodales de lenguaje grande para el control de robots, pero las capacidades incorporadas de los modelos siguen estando insuficientemente caracterizadas

La noticia

arXiv:2610.27865v1 Announce Type: new Abstract: Recent studies have explored the potential of multimodal large language models for robot control, yet the models' embodied capabilities remain insufficiently characterized across models, tasks, and execution settings. We conduct a multidimensional comparative evaluation on RoboColiseum across four tracks: Instruction Following, Spatial Reasoning, General Manipulation, and Robustness. We implement Agent2Act, a control harness that integrates GPT-6 Astra, GPT-6.1 Sol, Claude Opus 5.5, and Gemini 4 Argon into the evaluation environment through shared observation, action execution, and feedback interfaces. Under common task settings and scoring criteria, all four models exhibit a consistent capability profile: high instruction-following scores, with three approaching perfect performance; strong spatial reasoning; and