OpenAI's latest ChatGPT-4o version has surpassed Google Gemini, reclaiming first place on the LMSys Chatbot Arena leader-board. The new ChatGPT-4o, which was published on 8 August, 2024, scored 1314 points, outscoring Gemini-1.5-Pro-Exp by 17. This development follows Google's recent proclamation of its leadership position during the Made by Google speech.
The most recent version of ChatGPT-4o, which was released just a day before the Chatbot Arena results were declared, represents considerable improvements over its predecessors. The model showed benefits in technical domains, including coding, instruction following and managing complex prompts. According to LMSys.org on X, ChatGPT-4o surpassed its predecessor by more than 30 points in coding capabilities.
The new model's improved performance was visible in a variety of tests, including a significant accomplishment in which it built a full iOS app in just one hour. This efficiency, combined with previous updates to the Mac program, represents a huge step forward for ChatGPT users.
In contrast, Google's Gemini-1.5-Pro-Exp, which was formerly dominant, is now trailing. Google recently boasted about its leadership position during its presentation, but the abrupt drop in rankings highlights the quickly developing nature of AI technology.
The competitive scenario is further exacerbated by the upcoming releases of Google Ultra 1.5 and Claude Opus 1.5, as well as xAI's Grok 2, which has recently joined the top 10 ranks.