OpenAI has released a new update for the GPT-4o model. The update is said to improve the model’s creative writing capabilities and make it more engaging and natural. Moreover, it improves the model’s handling of uploaded files.
“It’s also better at working with uploaded files, providing deeper insights & more thorough responses,” said OpenAI on X.
The update is titled ‘gpt-4o-2024-11-20’, as mentioned by OpenAI Developers, on X.
A few social media users were quick to test out the update. One user on X mentioned that the model bypassed some of the most popular AI text detector tools.
“The GPT-4o update that dropped today fools most "AI detectors" and has a qualitatively different style than all previous models,” said the user.
Moreover, the model was also benchmarked on lmarena.ai, a benchmark to compare and evaluate LLMs through live, community-driven assessments. The latest GPT-4o ranked first, with a new overall high score on the leaderboard.
Exciting News from Chatbot Arena❤️🔥
— lmarena.ai (formerly lmsys.org) (@lmarena_ai) November 20, 2024
Over the past week, the latest @OpenAI ChatGPT-4o (20241120) competed anonymously as "anonymous-chatbot", gathering 8,000+ community votes.
The result? OpenAI reclaims the #1 spot, surpassing Gemini-Exp-1114 with an impressive 1361 score!… https://t.co/Q7q3Uonp94 pic.twitter.com/ogmhhCW7zY
On another benchmark by Aider, which evaluates an LLM’s code editing skill, it was said that the update to GPT-4o may be the first one that wasn’t worse than its predecessor.
GPT-4o, with a 128k context window, is currently the most advanced model available on OpenAI’s portfolio. As per OpenAI, it is the best model for ‘complex, and multi-step’ tasks and is cheaper, and more capable than GPT 4 Turbo.
And that’s not all. OpenAI also made the Advanced Voice Model available on desktop for all paid users and no longer exclusive to the mobile app. In September, the feature was rolled out to all ChatGPT Plus and Team users on the mobile app.
That said, we’re still waiting for the release of o1. The competition is also heating up, with the release of a new powerful reasoning model from DeepSeek, which offers a comparable performance to the o1 preview in math, and reasoning.
So when is the highly, anticipated release date? Around the corner, yet No1 knows!
Your reaction
Discussion
What's your take on this story?
No comments yet — be the first to share your view.