ChatGPT leads the way
✴️ChatGPT leads the way ✴️
Who can test models on the platform?
π The Allen Institute has launched the SciArena platform for testing AI models in the scientific field. Unlike Chatbot Arena, only scientists with publications who have undergone training participate here. The system produces two answers from random models based on articles from Semantic Scholar, and the researcher chooses the best one. The leader was ChatGPT o3 (1172 points), ahead of Claude Opus 4 and Gemini 2.5 Pro. It leads in all key disciplines. The platform excludes unreliable sources, strictly controlling the selection of literature - this can help improve scientific AI in the future.π―
#Chatgpt
Comments
Post a Comment