Models are stochastic. Did you look at pass@k? I wouldn’t be surprised if you saw a regression because these models are extremely complex and impact of various decision making downstream is complex.
replyI ran this multiple times through GPT-4 and every single time it arrived at the same conclusion. The data was readily available and pretty clear. GPT-5 insisted that the objectively inferior option was better until I gave it my own benchmark data and it was like "Oh okay nevermind".
replyGemini's answer was very opinionated and factually correct, whereas Claude gave a more nuanced answer, which was also very good.
This sounds perfectly reasonable and consistent with our current understanding of these models
reply