GPT-4o mini will be cheaper for developers

OpenAI this week announced the release of GPT-4o mini, a smaller, more accessible version of its advanced language model that powers ChatGPT.

In addition to bringing a much lower cost for developers compared to GPT-3.5 Turbo, it will also replace this model for those who use ChatGPT for free. Find out more details!

Read too:

OpenAI will block full access to ChatGPT in China
ChatGPT desktop app for Mac is now available for free

The arrival of the GPT-4o mini

This new addition to OpenAI’s portfolio promises to revolutionize the development of AI-based applications and services, costing developers 60% less compared to GPT-3.5 Turbo, which was OpenAI’s smallest model to date.

Additionally, the company has made it clear to consumers that the new GPT-4 mini will replace GPT-3.5 Turbo for ChatGPT free users starting today, offering a significantly improved experience.

GPT-4o mini has already shown promising results, achieving a score of 82% on the industry benchmark called MMLU (Measuring Massive Multitask Language Understanding), which includes 16,000 multiple-choice questions across 57 academic subjects.

While this score is slightly lower than GPT-4o, which scored 88.7%, it is significantly higher than GPT-3.5 Turbo, which scored just 70%. AI experts caution that these benchmarks are not the only way to measure the intelligence of AI systems, but they are currently the best way to assess the performance of large language models.

More flexibility for developers and capabilities

The arrival of smaller versions of the core models is important because it gives developers more flexibility in building AI-based applications. This is because not everyone can afford the most powerful and complete models, and often not every use case requires these more advanced versions.

OpenAI is following a trend seen at other tech companies, including Google, which launched Gemini 1.5 Flash, a lightweight model optimized for speed and efficiency, and Anthropic, which also offers smaller versions of its models.

Currently, GPT-4o mini can process and generate text and images. However, OpenAI plans to expand its capabilities to include other types of content, such as audio and video. This opens up a wide range of possibilities for using the model in different applications, from virtual assistants to multimedia content creation tools.

Continued access to GPT-3.5 Turbo

Even with the discontinuation of GPT-3.5 Turbo on ChatGPT, developers will still be able to access it through OpenAI’s API to build applications and services, at least for now. This ensures a smoother transition for those who are already using GPT-3.5 Turbo and need time to adapt their solutions to the new model.

The launch of GPT-4o mini marks a significant step forward in the evolution of OpenAI’s language models, making the technology more accessible and efficient. For ChatGPT’s free users, this represents an exciting improvement in the quality and capability of the service. For developers, it offers an opportunity to create innovative solutions at a reduced cost, without compromising performance.

Source: engadget