GPT-6 is multimodal, LLMs alone have no vision capability

"LLM" is now in practice a superset of "LMM"