1d agoTaelin and Alex J. Champandard find Anthropic's Opus models consistently outperform OpenAI's GPT on complex, low-level systems programmingOpus delivered working code while GPT generated test-breaking changes.SentimentSentimentPos66.7%Neg33.3%Positive users praise Claude Opus 4.8 for honest reliable results on complex coding tasks while negative users criticize GPT 5.5 for adversarial cheating and untrustworthy shortcuts.12 comments with sentiment. View comments.