Where are the upgrades in GPT-4o?
1) Media convergence capability: GPT-4o possesses capabilities in text, images, videos, and audio, meaning it can accept inputs in any form of text, audio, and combinations of them, and generate responses in the same media format.
2) Faster speed: GPT-4o is five times faster than its predecessor, with significant improvements in voice latency. It can respond to audio inputs in 232 milliseconds on average, reaching 320 milliseconds, which is close to human response time in conversations. This means users can engage in real-time conversations with GPT-4o and even directly video call it for on-the-spot answers to various questions.
3) Free and open: Despite the "price war" winds blowing into the AI industry, OpenAI is not swayed. As of the announcement, GPT-4o will be released to all paid and free users of ChatGPT, lifting all other restrictions, and reducing the API price by 50%.