OpenAI CEO Sam Altman has acknowledged the need for a new naming system for ChatGPT and its various versions following the release of their latest model, the ChatGPT-4o Mini. Since its inception, OpenAI has consistently used the same naming conventions for its ChatGPT models.
On July 18, OpenAI announced the ChatGPT-4o Mini, which is described as their "most cost-efficient small model." Altman shared the announcement on his X account (formerly Twitter), highlighting its affordability: "15 cents per million input tokens, 60 cents per million output tokens, 82% MMLU, and fast." He emphasized that the new model would be well-received by users.
While many users praised the new product, one user noted the need to change the naming scheme for ChatGPT models. Responding to Altman's post, the user commented, "You guys need a naming scheme to revamp so bad." Altman replied humorously, "lol, yes we do," indicating that a change might be considered.
The GPT-4o Mini is one of OpenAI's most effective AI models, known for its small size and minimal latency. It supports text, image, video, audio processing, and a vision and text API. The model features a context window of 128K tokens, can handle up to 16K output tokens per request, and incorporates knowledge up to October 2023. The new tokenizer, shared with GPT-4o, enhances cost-effectiveness for handling non-English text.
OpenAI emphasized the model's safety, stating that it adhered to its Preparedness Framework and underwent both automatic and human evaluations. Seventy external experts from various fields also assessed the AI model to identify potential risks.