OpenAI cuts the cost of using its AI with ‘mini’ models

OpenAI today announced a cut-price “mini” model that it says will allow more companies and programs to take advantage of its artificial intelligenceThe company says the new model, called GPT-4o Mini and available starting today, is 60 percent cheaper than OpenAI’s cheapest existing model and offers higher performance.

OpenAI The move is described as part of an effort to make AI “as widely accessible as possible,” but it also reflects growing competition between AI cloud providers, as well as a growing interest in small and free open source AI models. Meta is expected to launch the largest version of its very capable free offering, Llama 3, next week.

“The whole purpose of OpenAI is to create and distribute AI safely and to make it widely accessible,” Olivier Godement, the OpenAI product manager responsible for the new model, told WIRED. “Making intelligence available at a low cost is one of the most efficient ways for us to do that.”

Godement says the company has developed an affordable offering by improving the model architecture and refining the training data and training regime. OpenAI says GPT-4o Mini outperforms other “small” models on the market in several common benchmarks.

OpenAI has gained a significant foothold in the cloud AI market thanks to the remarkable capabilities of its chatbots. ChatGPTwhich began in late 2022. The company gives outsiders access to the large language model that powers ChatGPT, called GPT-4o, for a fee. It also offers a less powerful model called GPT-3.5 Turbo for about a tenth of the cost of GPT-4o.

The tremendous success of ChatGPT has fueled interest in language models, prompting competitors to develop similar offerings. Google, a leader in AI, has made a major effort to build and commercialize a large language model and chatbot under the brand name Gemini. Startups such as Anthropic, Cohere, and AI21 have raised millions of dollars to develop their own large language models and market them to business customers and developers.

Building the highest-performing large language models requires massive financial resources, but some companies have opted to open source their creations to attract developers to their ecosystem. The most prominent open source AI model is Meta’s Llama; It can be downloaded and used for free, but Its license imposes some limitations on commercial use,

This April, Meta Llama 3 announcedIts most powerful free model is . The company has released a smaller version of the model with 8 billion parameters — a rough measure of the model’s portability and complexity — as well as a more powerful, medium-sized, 70-billion-parameter version. The medium-sized model comes close to OpenAI’s best offering on a number of benchmark scores.

Multiple sources have confirmed to WIRED that Meta plans to release the largest version of Llama 3 with 400 billion parameters on July 23, though they say the release date could change. It’s unclear how capable this version of Llama 3 will be, but a few companies have plans to release it. Turning our attention to open source AI models Because they are cheaper and more customizable, and provide more control over the model and the data fed into it.

Godement acknowledges that customer needs are changing. “What we are seeing in the market is that developers and businesses are combining small and large models to create the best product experience for themselves at a reasonable cost and latency,” he says.

OpenAI’s cloud offerings give customers models that have undergone more security testing than competitors, Godement says. He adds that OpenAI could eventually develop models that customers could run on their own devices. “If we see a huge demand, we could open that door,” he says.