What models are performing better?

Guys, what models are performing better for you? for me, GPT-5-low and Claude Sonnet 4