GPT-6.1 SOL vs Opus 5.5: same samurai prompt, two different animations
Given the same prompt, Opus 5.5's rooftop chase animation carries the dango through the whole scene and adds a comic ending with an orange cat eating it, while GPT-6.1 delivers a clean fight.
🚨 GPT-6.1 SOL vs Claude Opus 5.5. Same prompt, same anime scene, and one of them made me actually care about a stick of dango Two samurai. Three pieces of dango. One rooftop chase above the clouds. Both models got the exact same prompt. Top: GPT-6.1 SOL. Bottom: Claude Opus 5.5. GPT-6.1 gave me a clean fight. Nice swords, nice bridge, nothing wrong with it. Opus gave me a story. The dango actually travels through the whole chase, the camera follows it, the moon and the floating city sit behind the roofs. Then the ending: a fat orange cat quietly eats the whole plate while both fighters sit there like nothing happened. I watched the first half as a fight scene and the second half as a comedy. That's the difference. Which one would you pick? 👀