AI startup OpenAI introduced a new AI model called "GPT-4o mini" on Thursday, marking the company's latest effort to expand the usage of its popular chatbot.
This mini AI model is a derivative of GPT-4o, which is the fastest and most powerful model released by OpenAI, capable of processing audio, video, and text in 50 different languages.
The "o" in the model stands for omniscience, with OpenAI planning to integrate images, videos, and audio into it in the future.
According to OpenAI's public test data, GPT-4o mini has excellent reasoning abilities in math, coding, and vision, making it one of the strongest multi-modal small models currently available.
For instance, GPT-4o mini scored 82.0% on the MMLU, 87.0% on the MGSM test in math and coding, and 59.4% on the multi-modal MMMU, outperforming Gemini Flash and Anthropic's Claude Haiku.
OpenAI has grown with the support of investors holding over $80 billion, continuously striving to maintain its leading position in the generative AI market and explore profitable avenues.
Last year, OpenAI's Chief Operating Officer Brad Lightcap stated: "The world is multi-modal. If you consider how we, as humans, process and interact with the world, we see things, we hear things, we speak things — the world is far more than just text. So for us, having text and code as a single mode, a single interface for how powerful these models can be and what they can do, always felt incomplete."
OpenAI announced on Thursday that GPT-4o mini will be available to free users of ChatGPT, ChatGPT Plus, and Team subscribers, with ChatGPT Enterprise users gaining access next week.
Key Points:
🤖 OpenAI launches "GPT-4o mini" model, bringing new upgrades to ChatGPT
🌐 GPT-4o mini is a derivative of the GPT-4o model, capable of handling multiple languages' audio, video, and text
💰 OpenAI is committed to maintaining its leading position in the AI market and exploring profit models