8h agoOpus 4.7 And GPT-5.5 Reach Nearly 90% On Agentic WeirdMLSentimentSentimentPos100%Neg0%Users are excited about Agentic Claude Opus 4.7 and GPT-5.5 reaching 90% accuracy on the WeirdML Benchmark because it shows strong model performance and room for further testing improvements.2 comments with sentiment. View comments.