The ChatGPT 4o-mini model is receiving a major update: discover 4 of its best new features.
Small but powerful.
OpenAI has improved its mini GPT-4o model of ChatGPT, incorporating a set of new features that surprisingly have been presented without the usual enthusiasm. This smaller and faster model has lacked many of the functions of its larger version until now, although, by design, its ability to perform the same tasks is inferior. Nevertheless, the efficiency and speed of the mini model have already been attractive to many ChatGPT users. Below are four of the main updates that bring it closer to the performance of the GPT-4o model.
One of the most notable new features is the ability to create images from text prompts, similar to what the full GPT-4o model allows. While there is no functional difference in visual generation, the smaller size of the mini model could speed up the production process. However, the final result should not differ significantly, as both models use OpenAI's DALL-E 3 image generator.
Previously, one of the most notable limitations of the mini model was its lack of internet access. Now, users can utilize it through ChatGPT to conduct online research faster than with the GPT-4o. This allows for updated information, instead of relying solely on its pre-trained knowledge base. This improvement is particularly relevant for fact-checking and data comparison, thus elevating its functionality to match that of its larger model, though always tied to the speed of the mini model.
Another limitation of the mini GPT-4o was the lack of options for uploading documents. However, now users have the ability to analyze, summarize, and discuss documents and images that they upload. The visual processing features of the model allow it to understand and explain the uploaded images, just as it does with text. This functionality facilitates working with complex visual data, both for educational and personal purposes.
Finally, the mini model has incorporated a new memory feature that allows it to remember previous conversations with specific users. This means that even after a conversation ends, when returning to interact with the model, it will be able to provide more relevant responses and continue previous discussions, adapting to the user's preferences. In summary, being a mini model does not mean it cannot handle long-term interactions.
