3 ms·Companies are optimizing models for specific benchmarks4 points by mzubairtahir 3mo agoOpenai 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