Claude Opus 4.8 went live in DeepSWE and it’s rank 2# with 58% success Still 12% behind gpt 55 Only 4% better than 4.7 I don’t know how OpenAI get this crazy models

0 likes 0 replies