News

During its along with a and post, OpenAI has announced the launch of its (“o” for “omni”), that has near human response times.  OpenAI says the new model will be able to respond to audio inputs in as little as 232 milliseconds, which is similar to human response time in a conversation. While GPT-4o can receive queries in any combination of , , and and will generate responses in the same vein, it was demoed using voice mode to talk to ChatGPT. On the blog page, you can get one of six preset voices read the page to you. The voices included three male and three female voices. According to the demos, GPT-4o can recognise and respond to screenshots, videos, photos, documents, uploaded charts, and also recognise facial expressions, and handwritten information. Demos we found interesting was the ability to create a detailed summary of a video presentation and the ability to create a summary of a meeting with multiple attendees from a voice recording. Transcription tool could be in trouble here. Developers can also now access GPT-4o in the API as a text and vision model for half the price, and with 5x higher rate limits compared to GPT-4 Turbo. There will also be a new desktop app on offer for the Mac, with a Windows version coming soon.