Photo: Camilo Concha/Shutterstock.com
On Monday, May 13, 2024, OpenAI’s Spring Update event was live-streamed on YouTube. During the event, OpenAI announced the launch of their latest model, the GPT-4o ‘omnimodel’. This new version of the GPT-4 series offers significant improvements in speed and capabilities across text, vision, and audio processing.
OpenAI allows free users to access its GPT Store and all its features. Lastly, GPT also receives a desktop app for Mac.
GPT-4O
The core innovation of GPT-4o lies in its ability to democratise access to advanced AI intelligence, extending the level of sophistication previously reserved for premium users to all, including those utilising OpenAI’s services for free.
During a live stream presentation, OpenAI’s CTO Mira Murati highlighted the key features of GPT-4o, emphasising its enhanced ease of use and speed, positioning it as a significant leap forward in AI user experience.
The unveiling showcased GPT-4o’s real-time voice interaction capabilities within ChatGPT, showcasing faster responses, improved conversational skills, and the ability to integrate with users’ workflows.
The tool can provide real-time tutorials, AI-generated storytelling with diverse voice options, and complex problem-solving engagements, all of which demonstrate GPT-4o’s versatility and utility.
One of the standout features of GPT-4o is its multimodal capabilities, which combine visual, audio, and text interactions. This was exemplified by demonstrations such as language translations, emotion detection, and utilising a phone’s camera for tasks like note reading.
Looking ahead, OpenAI outlined a phased rollout plan for GPT-4o’s capabilities. Text and image functionalities are already available to paid users and will be extended to free users in the future. The voice version of GPT-4o is also scheduled for release shortly, further enhancing ChatGPT’s capabilities.
“GPT-4o (Omni) is much better than any existing model at understanding and discussing the images you share. For example, you can now take a picture of a menu in a different language and talk to GPT-4o to translate it, learn about the food’s history and significance, and get recommendations,” said the company.
GPT desktop app
OpenAI also announced a desktop app for ChatGPT with a refreshed user interface currently exclusive to MacOS users.
The demo presented the features of the ChatGPT desktop application, which enables users to open it in a compact window simultaneously with other software. Users can communicate with ChatGPT by asking questions about the content on their screen, either by typing or using voice commands. This AI-powered assistant can respond based on its comprehension of the visual context.
One notable feature is the Option + Space keyboard shortcut that lets users query ChatGPT quickly. Additionally, users can capture and discuss screenshots directly within the app, enhancing the collaborative experience.
Both free and paid ChatGPT users will gain access to the desktop app, starting with ChatGPT Plus subscribers and expanding to all users in the upcoming weeks. OpenAI plans to launch a Windows app later this year, extending its accessibility across different platforms.
Alongside the desktop announcement, OpenAI showcased a new UI for ChatGPT on the web with minor changes to the home screen and message layout.
GPT Store is now accessible to free users
In a move towards democratising access to advanced AI capabilities, OpenAI has announced that previously subscription-only features like the GPT Store of ChatGPT will be available to free users.
This will allow free users to browse and utilise custom bots created by ChatGPT Plus subscribers. OpenAI introduced the GPT Store to paid subscribers just four months ago, in January 2024, as a platform for creating and sharing custom chatbots.
While the engagement-based revenue-sharing program for GPT builders began testing in March, its full potential was limited by the feature’s restriction to paying ChatGPT users.
In the News: WhatsApp tests channel access to linked devices on iOS and Android