GPT-4o: the new version of ChatGPT that is faster and more powerful

A OpenAIone of the hottest companies at the moment, announced this Monday (13) the launch of ChatGPT for desktop environment and the official presentation of its new state-of-the-art language model (LLM), the GPT-4o (they say “four oh”).

The latter appears as interesting news for all users, as the new LLM will be offered free of charge to the general public, including those individuals who do not have a subscription to the ChatGPT Plus premium service. Therefore, free mode users will not be limited to version 3.5 only

Read too
Is it worth subscribing to ChatGPT Plus? Understand the differences to the free version
ChatGPT will be used by the government of the State of SP to generate teaching material for digital classes

It is important to note that, along with these expressive announcements, there was also an announcement related to the ChatGPT interface, which was renewed to provide an even more intuitive and responsive experience.

Most advanced OpenAI model

GPT-4o is the most advanced AI model ever released by OpenAI. Capable of responding to audio inputs in just 232 milliseconds – similar to the human response time in a conversation – the company promises several improvements for text, and also in the understanding of images and audio, when compared to previous models.

According to OpenAI CTO, Mira Murati, “This is the first time we’re really taking a big step forward when it comes to ease of use.”

The new model has the ability to provide responses in different tones of voice and can even detect users’ moods to enrich the conversation. According to the official announcement, GPT-4o is a step towards a much more natural human-computer interaction, and accepts as input any combination of text, audio and image, generating a combination of outputs in these formats.

Simultaneous translation

With ChatGPT powered by GPT-4o, OpenAI wants the chatbot to also be an ally when interacting in different languages. Example, let’s assume you are interacting with someone who speaks English, ChatGPT can, for example, translate the conversation in real time into Portuguese.

After the AI ​​understands that it is time to continue with the translation, just continue talking normally, taking advantage of simultaneous translation. It is also interesting to observe how technology can quickly shape its way of responding to a question based on user interruptions in search of a new pattern. Check it out below.

Real-time analysis

One of the strengths of GPT-4o is real-time multimodal analysis. Using the cell phone’s selfie camera, AI can analyze the user’s clothing, the environment they are in, or even check whether a mathematical problem is correct.

Just as is already possible with GPT-4, ChatGPT powered by GPT-4o will also be able to respond to the user based on an image and screenshot passed to them. You can send a photo and ask for information about what it is. Knowing which model a particular car is, for example. GPT-4o will perform this analysis with more precision and speed.

ChatGPT desktop for macOS only

The launch event, broadcast live on the internet, was attended by Mira Murati, the CTO of OpenAI, who took on the role of main presenter, masterfully conducting the announcements.

The company surprised many by announcing that the ChatGPT desktop application would initially be dedicated only to users of the macOS operating system. With this new version, users can activate and interact with the AI ​​through a simple and agile shortcut, pressing the Space + Option keys simultaneously.

In technical terms, OpenAI shared that GPT-4o boasts significant improvements over its previous version, called GPT-4 Turbo. Not only is it faster at processing information, it is also more energy efficient. The suffix “o” placed in the name of the new LLM is an allusion to the word “omni”, derived from Latin, which means “everything”. This lexical choice symbolizes the comprehensive and ubiquitous character of this new LLM, which demonstrates notable advances in both speech recognition and processing and image analysis.

GPT-4o will be available for free

With GPT-4o, OpenAI establishes this model as the best version available for ChatGPT’s free mode, displacing GPT-3.5, available since 2022.

Starting today, it will be possible to access the new feature related to text and image analysis – a feature that was only available in paid mode. The release is occurring gradually.

GPT-Plus subscribers will also have some advantages, such as accessing personalized GPTs, based on specific data training, such as a chatbot aimed at analyzing scientific texts. The new voice mode integrated into ChatGPT is also reserved for subscribers. Subscription costs US$20 (R$102.81).

The free version also has a usage limit. ChatGPT Plus subscribers will have a limit of up to 5 times higher than the free version. Team and Enterprise subscribers will have even more capacity.