“Ethan Mollick asserts that both GPT-5.6 SOL and Fable-5 represent significant advancements over previous models, creating a large performance gap with the next-best AIs.”