Companies are optimizing models for specific benchmarks4 points·mzubairtahir··0 commentsOpenai 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 examSaveView on HN