The landscape of AI is shifting. OpenAI is set to retire GPT-4, a model that has been a significant part of ChatGPT since its launch over two years ago. This change is expected to happen by April 30. The new model, GPT-4o, will take over as the primary model in ChatGPT. GPT-4 will still be accessible through OpenAI’s API, though.

GPT-4o has shown impressive capabilities. It outperforms GPT-4 in various tasks, such as writing, coding, and solving problems in science, technology, engineering, and mathematics. Recent updates have enhanced GPT-4o’s ability to follow instructions, solve problems, and engage in conversation. This makes it a strong candidate to replace GPT-4.

GPT-4 made its debut in March 2023. It was integrated into ChatGPT and Microsoft’s Copilot chatbot. Some versions of GPT-4 could handle both images and text, a first for OpenAI’s widely used models. The cost to train GPT-4 was substantial, reportedly over $100 million. This investment led to the development of GPT-4 Turbo in November 2023, which is faster and more cost-effective.

The retirement of GPT-4 comes amid legal disputes. Publishers, including The New York Times, have accused OpenAI of using their data to train GPT-4 without permission. OpenAI argues that fair use laws protect them from these claims. The outcome of these disputes could shape the future of AI development and data usage.

OpenAI is not stopping with GPT-4o. The company is preparing to launch new models. These include a family of models called GPT-4. 1, which includes GPT-4. 1-mini and GPT-4. 1-nano, as well as reasoning models o3 and o4-mini. These new models promise to push the boundaries of what AI can do.

The transition from GPT-4 to GPT-4o marks a significant moment in AI evolution. It highlights the rapid pace of development and the constant quest for improvement. As AI models become more advanced, they raise important questions about data usage, copyright, and the future of technology.

OpenAI's Next Step: GPT-4's Exit and GPT-4o's Rise

Actions