Four AI models generate the same arctic tugboat scene: cost vs detail
Given one prompt, Claude Opus 5.5 spent $45.97 and 3h 11m for the most detailed shot, while cheaper GPT-6 variants delivered less, leaving the question of whether 30x the price buys 30x the picture.
I gave 4 AIs the same prompt: an arctic night, a tugboat, a whale, an iceberg arch🚢🐋 Claude Opus 5.5: $45.97 · 3h 11m GPT-6.1 Sol: $3.11 · 54m GPT-6 Astra: $17.51 · 1h 00m GPT-6.1 Terra: $1.49 · 59m Opus cost 30x more than Terra But is it 30x better? Look closer Claude Opus 5.5: the most detail. The ship is named “NORDLYS”, crew stand on deck, the whale spouts, the hull reflects in the water, and aurora glows over snowy peaks. This is the one that looks like a film shot GPT-6.1 Sol: the best value. A warm lit boat, crew on deck, a whale beside it and a clean wake, for $3.11. GPT-6 Astra: the most dramatic ice wall, but at $17.51 it’s hard to say why it costs 5x more than Sol. GPT-6.1 Terra: the cheapest, but a tiny boat on flat dark water. It works, but it feels empty. Nobody sweeps Opus wins on detail. Sol wins on value. Terra wins on price. So which one are you actually paying for, the best picture or the best deal?