Is it safe to say kimi k3 can beat the opus 4.7 and gpt 5.4 ? (In agentic task and frontend and backend coding.) (All at max mode)

Well opus 4.8 and gpt 5.5 was released recently so i think its better to take the older model. In the case, if it cannot beat any one of them, in any particular task, then is it safe to say that it is at least beating the Opus 4.6 in most tasks? :( Well I know the Kimi 3 is still not open sourced, b