Companies are optimizing models for specific benchmarks

By mzubairtahir · 2026-07-25 · 2 points · 0 comments

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