OpenAI updated ChatGPT and regained first place in the LLM leaderboard

OpenAI just made it ChatGPT a much better writer. Latest update for basic model GPT-4o gave AI a boost in creativity.

This new update also helped leapfrog Google Gemini and regain first place in the LLM arena. This is a virtual platform where users evaluate the performance of two models without fully knowing which one.

OpenAI didn’t say much about the new update other than to say that its “creative writing abilities have been enhanced” and that it is now “more natural and engaging” with “more tailored writing to improve relevance and readability.”

What’s new in ChatGPT?

Exciting news from the chatbot arenaOver the past week, the latest @OpenAI ChatGPT-4o (20241120) anonymously competed as an “anonymous chatbot” with over 8,000 community votes. Result? OpenAI regains first place, overtaking Gemini-Exp-1114 with an impressive score of 1361 points!… pic.twitter.com/ogmhhCW7zYNovember 20, 2024

OpenAI recently kicked off its game by adding a number of new features, including bringing Advanced Voice to the Internet and hints at best generations of DALL-E images. There are even rumors of an imminent release of a version of Sora that will bring AI video to the rapidly evolving AI platform.

I tried it and it seems more complete and interesting in its answers.

The latest update is more of a “behind the scenes” update than a new UI or flashy new features like impressive canvas. This is a change in how GPT-4o works that makes it a more powerful creative tool than previous generations.

The new version of GPT-4o, still called GPT-4o, also works better with files you upload to ChatGPT and provides deeper insight into their contents.

It was secretly released on the LLM chatbot arena lmarena.ai (formerly LMSys arena). It is a platform where models anonymously compete to see which ones are rated better by human users.

Last week OpenAI unveiled it as an “anonymous chatbot” and it quickly outperformed Gemini-Ex-1114, the latest model from Google who held the crown for about a week. In her article about X, Lmarena described it as a “remarkable improvement” over previous versions, including in the areas of creative writing, coding and mathematics.

Have you noticed changes in ChatGPT over the past few days? I tried it and it seems more complete and interesting in its answers. This increase in creativity may also explain why DALL-E’s images improved.