Кто пишет код лучше всех? Сравнил GPT‑5.2, Opus 4.5, Sonnet 4.5, Gemini 3, Qwen 3 Max, Kimi, GLM

In the new video, Oleg Stepanov conducts a comparison of top neural networks for programming, including models like GPT 5.2, Opus 4.5, Sonnet 4.5, and others. He begins with a brief overview of the models, noting that many new neural networks have recently emerged, setting records across various metrics. Oleg explains the approach he uses to test the models, including details about implementation and development environment. The first tests involve creating a simple fantasy-style 'to do list'. Oleg shares the results obtained, highlighting how each model handles the tasks, noting the shortcomings and strengths of each. Looking at the results, he discusses the issues encountered by certain models and points out which ones provided the most distorted results. The subsequent tasks include more complex systems, such as gathering information from Reddit and developing a custom admin panel for servers. Oleg emphasizes which models performed best and shares their features and utility in real-world scenarios. It is evident that GPT 5.2 stands out among other models due to its powerful functionality. In conclusion, Oleg shares his insights on which models performed the best and what is necessary for successful development using AI. He emphasizes the importance of studying documentation and testing in the process of project creation with neural networks. This video serves as an excellent source of information for developers interested in utilizing AI in programming.