Imagine having four of the leading language models in the world competing against each other in a live coding challenge. Sounds like a fascinating spectacle, right? That’s exactly what’s happening today. In a unique experiment, GPT-5, Claude Sonnet 4, Gemini 2.5 Pro, and GLM45 are being put to the test to see how they handle coding tasks in a natural-language coding environment.
The models will receive the same prompt and will be tested on different complexity levels, including simple builds, bug-fix tasks, multi-step complex builds, and possible planning flows. The goal is to compare their output quality, build speed, and debugging performance.
What’s even more interesting is that you can join in on the action and ask questions, suggest prompts, or simply observe how these language models work their magic. The test is happening live, so don’t miss out on the opportunity to see the future of coding unfold.
Will GPT-5 emerge as the winner, or will one of the other models surprise us with its capabilities? Tune in to find out.