🧵 4/ Despite strong results, '01' model shows less progress in creative writing tasks compared to GPT-4o. Demonstrates chain-of-thought AI process for improved outcomes.