Google and OpenAI have taken turns battling for the top spot on LMSys’s Chatbot Arena rankings. The past week has been particularly intense as they traded victories multiple times within the same week.
This round of competition kicked off with Google releasing the exclusive Gemini-Exp-1114 API, available only on AI Studio. This model defeated OpenAI’s pricey models, o1-preview and o1-mini. However, OpenAI then released the ChatGPT-4o-latest model on November 20, reclaiming victory over Gemini once again. Google then released Gemini-Exp-1121, scoring another win against the latest iteration of ChatGPT-4o.
Since the models were recently released for testing, Gemini-Exp-1121 received fewer than five thousand votes, with a wide 95% confidence interval of +7/-9. Despite this, the overall leaderboard is currently dominated by the intense rivalry between Google and OpenAI.
Other interesting models in the top 10 include Grok-2-08-13, alongside Yi-Lightning, GPT-4o, and Claude 3.5 Sonnet.
Source: LM Arena
TLDR: Google and OpenAI battled fiercely in the Chatbot Arena rankings, trading wins multiple times, with Google’s Gemini and OpenAI’s ChatGPT models taking center stage.
Leave a Comment