Today, OpenAI announced a new update for the GPT-4o model, which powers the popular ChatGPT AI assistant. This latest LLM update introduces several improvements over the existing GPT-4o model, released on January 29th.
OpenAI stated that this improved model is better at following detailed instructions, particularly prompts containing multiple requests. It has also enhanced its ability to tackle complex technical and coding problems. Additionally, the model delivers responses with improved intuition, creativity, and fewer emojis.
From ChatGPT Release Notes:
Smarter problem-solving in STEM and coding:
GPT-4o has further improved its capability to tackle complex technical and coding problems. It now generates cleaner, simpler frontend code, more accurately thinks through existing code to identify necessary changes, and consistently produces coding outputs that successfully compile and run, streamlining your coding workflows.
Enhanced instruction-following and formatting accuracy:
GPT-4o is now more adept at following detailed instructions, especially for prompts containing multiple or complex requests. It improves on generating outputs according to the format requested and achieves higher accuracy in classification tasks.
“Fuzzy” improvements:
Early testers say that the model seems to better understand the implied intent behind their prompts, especially when it comes to creative and collaborative tasks. It’s also slightly more concise and clear, using fewer markdown hierarchies and emojis for responses that are easier to read, less cluttered, and more focused. We're curious to see if our users also find this to be the case.
OpenAI CEO Sam Altman tweeted that the improvements in this update are substantial and that more updates are coming soon.
we updated GPT-4o! really big strides.
— Sam Altman (@sama) March 27, 2025
more updates to come. https://t.co/gCv6d1jk65
The updated GPT-4o model is now available to all ChatGPT paid users. Free users will receive this model update over the next few weeks. For developers, the new model is accessible via the chatgpt-4o-latest API. OpenAI also plans to extend these improvements to an older model in the API in the coming weeks.
On the LMArena crowdsourced AI benchmarking site, the latest ChatGPT GPT-4o (2025-03-26) model has reached the No. 2 position, even surpassing its own GPT-4.5, which was launched last month. Compared to ChatGPT's GPT-4o (2025-01-29), this updated model scored 30 points higher.
In key categories where it previously lagged, it has made significant improvements. In the Mathematics category, it improved its ranking from No. 14 to No. 1. In Hard Prompts, it is now ranked No. 1, a major leap from No. 7. In Coding, it has also secured the No. 1 spot, a significant jump from No. 5.
As OpenAI continues to refine and improve its model capabilities, ChatGPT users can look forward to even more improvements in the near future.
5 Comments - Add comment