OpenAI benchmark charts become a meme after o1, with GPT-5 and o3 graphics mocked

willdepue · x · 2026-07-22

The post mocks OpenAI’s design team for allegedly “going off the rails” after o1 and making charts like the ones shown in the attached images.

The images themselves are polished benchmark-style graphics comparing GPT-5, o3, and GPT-4o on software engineering and “Humanity’s Last Exam,” making the whole post read like a parody of OpenAI’s presentation style.

Original post →

More from Fun

Fun channel →