OpenAI unveiled its latest model, the ChatGPT-4o Mini, hailed as "our most cost-efficient small model." Sam Altman- OpenAI’s Chief has announced the release on his X account (formerly known as Twitter) and highlighted the model's cost-efficiency, by writing: "15 cents per million input tokens, 60 cents per million output tokens, 82% MMLU, and fast."?
Altman further added, "Most importantly, we think people will really, really like using the new model."
Naming scheme revamp
While many users have appreciated the new model, there was one user who commented on the need for a change in the naming scheme for ChatGPT models. The comment read, "You guys need a naming scheme revamp so bad."?
Sam Altman acknowledged this feedback, responding with, "lol yes we do."?
This exchange also indicated that OpenAI was considering a new naming strategy for future ChatGPT versions.
ChatGPT-4o Mini: Features
The ChatGPT-4o Mini is said to be one of the most effective AI models from OpenAI, which has been designed to provide minimal latency and high efficiency.
It supported various formats, including text, video, image and audio, along with enhanced vision (image processing) capabilities.?
The model further features a context window of 128K tokens and supports up to 16K output tokens per request. It also incorporates knowledge up to October 2023 and boasts improved handling of non-English text, thanks to a new tokenizer shared with GPT-4o.
Safety and preparedness measures
In line with OpenAI's Preparedness Framework, the ChatGPT-4o Mini has gone through rigorous safety evaluations, both automated and human.?
OpenAI has collaborated with 70 external experts from various disciplines, to identify potential hazards. These measures have ensured that the AI model adheres to high safety standards and is ready for widespread use.
This approach has highlighted the commitment of OpenAI towards continuous improvement and user satisfaction.