"Drawing" the Mona Lisa with GPT-5.6, Claude, Gemini, and Grok

Editor's take

Large language models are now capable of generating visual art based on textual prompts, as demonstrated by a recent comparison of GPT-4o, Claude 3, Gemini 1.5 Pro, and Grok 1.5's attempts to recreate the Mona Lisa.

This development signifies a growing convergence of language and image generation capabilities within single AI models, moving beyond purely text-based tasks. The implications are significant for creative industries, potentially enabling new forms of digital art creation and content generation, while also raising questions about authorship and originality in a landscape increasingly populated by AI-produced visuals.

Future developments to monitor include the fidelity of these generated images compared to human-created art, the evolution of prompt engineering for nuanced artistic control, and the legal frameworks that will govern AI-generated artwork. The ability to consistently produce recognizable, aesthetically pleasing images across diverse styles will be a key indicator of progress.