公司正在针对特定基准优化模型。1 分•作者: mzubairtahir•3 个月前•原帖OpenAI正在优化GPQA钻石，而Anthropic则在优化人类最后的考试。GPT 5.6在GPQA中获胜，而Opus 5在人类最后的考试中获胜。查看原文Openai is optimizing for gpqa diamond and anthropic is optimizing for humanity last exam. gpt 5.6 wins on gpqa and opus 5 wins on humanity last exam
