Four AI Models Attempt to Draw the Mona Lisa and Other Artworks

Jul 21, 2026

AI Summary

Four AI models, including GPT-5.6 and Claude Fable 5, were tested on their ability to draw the Mona Lisa and Van Gogh's Starry Night using a colored-pencil toolset. The study evaluated their drawing techniques, costs, and output quality, revealing varied performance and efficiency among the models.

Four AI models were tested: GPT-5.6 Sol, Claude Fable 5, Grok 4.5, and Gemini 3.6 Flash.
Each model was given a blank canvas and a set of colored pencils to create drawings based on two target images and five open-ended prompts, resulting in 28 total drawings.
The models had access to the target images throughout the drawing process and were scored on structural similarity (SSIM) to the targets.
Claude Fable 5 had the highest cost at approximately $160 for seven drawings, while GPT-5.6 Sol was the least expensive at $7.74.
The models exhibited different drawing strategies; for instance, GPT-5.6 Sol did not use certain tool settings, while Grok 4.5 relied heavily on them.
Gemini 3.6 Flash achieved the highest peak SSIM score of 0.449 for the Mona Lisa but ended lower at 0.337, indicating that more reviews did not necessarily lead to better final outputs.
GPT-5.6 Sol was noted for its detailed drawings, outperforming others in quality, while Grok 4.5 produced less usable results.
Claude Fable 5 was ranked second in quality but was significantly more expensive and slower than the other models.
All models are available for use on a pay-as-you-go basis without subscriptions.

ai drawinggpt-5.6claudegeminigrok

Four AI Models Attempt to Draw the Mona Lisa and Other Artworks

Related Stories

AI Helps Saxboy Billy Rise to Global Fame in 2023

ChatTJB offers human responses to user queries without AI technology

Anthropic outlines marking system for AI-generated content under EU regulations

Chinese tech companies halt AI companion services, leaving users heartbroken