I don’t really like GPT-5.5…

A hands-on critical review of GPT-5.5, OpenAI's latest model. Despite being the most capable model yet with impressive benchmark scores and strong coding output, the reviewer finds it frustrating to use in practice. Key complaints include the model feeling 'lazy' by only barely honoring intent, difficulty recovering from incorrect context in long threads, and a significant price hike ($5/M tokens in, $30/M out). Token efficiency improvements are praised, and the Pro model is highlighted as genuinely impressive for hard problem-solving tasks like cipher puzzles. The reviewer recommends more upfront prompt detail, frequent new threads, and pre-researching context before starting tasks.